[subexp-daq] New warning meassage in R3BFUSER

Hans Toshihide Törnqvist hans.tornqvist at chalmers.se
Mon Jul 29 12:36:59 CEST 2024 

Dear all,

Looks like I went on vacation just before these troubles... I have not 
been able to give these codes the time they require to be in good shape, 
apologies.

The proposed solutions were fine and are online now.

Best regards,
Hans

On 2024-07-19 12:22, Håkan T Johansson wrote:
> 
> Dear Günter,
> 
> that sounds great!  Let's check that I got it right:
> 
> - You are now running with up-to-date (master/main):
> 
>    drasi,
>    trloii,
>    r3bfuser and
> 
>    nurdlib is at the 'newest' SIS3316_check_hit_reworked_rebased (ccac28b)
> 
> With the following fixes, it runs:
> 
> - r3bfuser: log_callback - second argument (line number) not 'int'
>    (I've sent a merge request)
> 
> - nurdlib: FUNC_NONNULL((1,2))
>    (I've sent a merge request)
> 
> - r3bfuser: changed:
> 
>          /* diff = crate_counter_get_diff(g_counter_ms); */
>          diff = 1;
> 
>    This will need some deeper thought to figure what really should be fixed
>    though.
> 
> Cheers,
> Håkan
> 
> 
> 
> 
> On Fri, 19 Jul 2024, Weber, Guenter Dr. wrote:
> 
>> Dear Håkan,
>>
>>
>> I now started with the software package from ~two months ago and 
>> step-by-step replaced it with the most recent versions. Everything is
>> fine until I add the new version of NURDLIB (using the 
>> SIS3316_check_hit_reworked_rebased branch). Then I get in R3BFUSER the 
>> compile
>> error.
>>
>>
>> Implementing the first fixes suggested by you elimates this error on 
>> the RIO4:
>>
>>
>> it looks like with a change of the second parameter of log_callback (both
>> in prototype an definition) of r3bfuser/f_user.c from unsigned to int,
>>
>>
>> But on startup of the DAQ I get a crash. Then we have the second fix:
>>
>>
>> Anyhow, I suspect that if you change in
>> include/nurdlib/crate.h
>> in the prototype of crate_dt_release_set_func to have
>> FUNC_NONNULL((1,2))  instead of  FUNC_NONNULL(())
>> then it might work.
>>
>> This still results in a crash of DAQ, as g_counter_ms in f_user.c is 
>> zero at line 981. The immediate crash of DAQ on startup is avoided
>> by a workaround:
>>
>>
>>         /* diff = crate_counter_get_diff(g_counter_ms); */
>>         diff = 1;
>>
>> And, (!) contrary to what I thought to having found yesterday (!), 
>> with this modifications the DAQ is now running as expected also with
>> a SIS3316 module being enabled.
>>
>>
>>
>>
>> Thus, it looks like that all problems seem to originate from a change 
>> in NURDLIB.
>>
>>
>>
>> I hope this helps with fixing the underlying issues.
>>
>>
>>
>>
>> Best greetings
>>
>> Günter
>>
>>
>>
>> _______________________________________________________________________________________________________________________________________
>> Von: subexp-daq <subexp-daq-bounces at lists.chalmers.se> im Auftrag von 
>> Håkan T Johansson <f96hajo at chalmers.se>
>> Gesendet: Donnerstag, 18. Juli 2024 17:39:43
>> An: Discuss use of Nurdlib, TRLO II, drasi and UCESB.
>> Betreff: Re: [subexp-daq] New warning meassage in R3BFUSER
>>
>> Dear Günter,
>>
>> thanks!  Yes, what I was looking for.
>>
>> r3bfuser - same as new version  (so should not be at fault ;) )
>>
>> Good news (to me) is that the nurdlib is in the same repo.
>>
>> There are changes in sis3316-code in nurdlib between the working c36ae0a
>> and non-working ccac28b which are beyond my understanding.  Unless urgent
>> for you, I would suggest that we wait with that until Hans is back from
>> vacation.  ;)
>>
>> Meanwhile, if you like, you could of course start out at the working
>> version combination, and update drasi and trloii to current version.  One
>> at a time.  The interfaces between those and also those and nurdlib are
>> small and should not have seen changes lately, so I would expect any such
>> combinations to compile.
>>
>> Cheers,
>> Håkan
>>
>>
>> On Thu, 18 Jul 2024, Weber, Guenter Dr. wrote:
>>
>> >
>> > Here are the hashes of the (old) software combination that runs 
>> without issues:
>> >
>> >
>> > (base) mbsdaq at atpnbg012:~/mbsrun/rio4/2024_mcalstruck/nurdlib$ git 
>> rev-parse --short HEAD
>> > c36ae0a
>> >
>> > (base) mbsdaq at atpnbg012:~/mbsrun/rio4/2024_mcalstruck/trloii$ git 
>> rev-parse --short HEAD
>> > 0020559
>> >
>> > (base) mbsdaq at atpnbg012:~/mbsrun/rio4/2024_mcalstruck/drasi$ git 
>> rev-parse --short HEAD
>> > 1c08b62
>> >
>> > (base) mbsdaq at atpnbg012:~/mbsrun/rio4/2024_mcalstruck/r3bfuser$ git 
>> rev-parse --short HEAD
>> > d09918b
>> >
>> > I hope this is what you were asking for.
>> >
>> >
>> >
>> > Best greetings
>> > Günter
>> >
>> >______________________________________________________________________________________________________________________________________
>> _
>> > Von: subexp-daq <subexp-daq-bounces at lists.chalmers.se> im Auftrag 
>> von Håkan T Johansson <f96hajo at chalmers.se>
>> > Gesendet: Donnerstag, 18. Juli 2024 16:50:42
>> > An: Discuss use of Nurdlib, TRLO II, drasi and UCESB.
>> > Betreff: Re: [subexp-daq] New warning meassage in R3BFUSER
>> >
>> > On Thu, 18 Jul 2024, Weber, Guenter Dr. wrote:
>> >
>> > >
>> > > Dear Håkan,
>> > >
>> > >
>> > > we have to DAQ systems with identical hardware. Whenever, there is 
>> a new version (TRLOII,
>> > > NURDLIB, DRASI, etc.) from your side that we want to test, I am 
>> setting up one of them
>> > > with the most recent software version.
>> > >
>> > >
>> > > So, right now we have a running system that is roughly on the 
>> software status of two
>> > > months ago (if I remember corretly, this was after the repository 
>> change as we first had
>> > > to ask you to recreate the SIS3316 development branch) with some 
>> modifications from our
>> > > side to the SIS3316 code (committed about six weeks ago).
>> >
>> > Just so I follow correctly: both systems are tested able to run with 
>> some
>> > (specific) software version?
>> >
>> > Could you send the git hashes of nurdlib/r3bfuser/trloii/drasi of that
>> > working combination?
>> >
>> > > And we have the 'new' system
>> > > that has identical hardware and the software status from last week.
>> >
>> >
>> > Cheers,
>> > Håkan
>> >
>> >
>> >
>> > >
>> > >
>> > >
>> > > Best greetings
>> > >
>> > > Günter
>> > >
>> > >
>> > >
>> > >
>> > > 
>> _________________________________________________________________________________________
>> > > Von: subexp-daq <subexp-daq-bounces at lists.chalmers.se> im Auftrag 
>> von Håkan T Johansson
>> > > <f96hajo at chalmers.se>
>> > > Gesendet: Donnerstag, 18. Juli 2024 15:47:46
>> > > An: Discuss use of Nurdlib, TRLO II, drasi and UCESB.
>> > > Betreff: Re: [subexp-daq] New warning meassage in R3BFUSER
>> > >
>> > > Dear Günter,
>> > >
>> > > did I follow this thread correctly that these tests are with 'new'
>> > > hardware, i.e. a setup which has no working reference software?
>> > >
>> > > You have / are using the SIS3316 modules in another physical setup 
>> with
>> > > some fairly recent branch in the previous nurdlib repository?
>> > >
>> > > https://gitlab.com/chalmers-subexp/nurdlib.prelgpl21/-/network/master
>> > >
>> > > --
>> > >
>> > > If so, there are three changes at the same time:
>> > >
>> > > - Nurdlib updates (also split by the repository change).
>> > >
>> > > - The SIS3316 code branch adapted from the previous to the new 
>> repository.
>> > >
>> > > - A new hardware setup.
>> > >
>> > > Cheers,
>> > > Håkan
>> > >
>> > >
>> > >
>> > > On Thu, 18 Jul 2024, Weber, Guenter Dr. wrote:
>> > >
>> > > >
>> > > > Dear Håkan,
>> > > >
>> > > >
>> > > > with diff = 1 the DAQ is running as long as only the VULOM4B is 
>> readout. Adding a
>> > > SIS3316
>> > > > module results in an readout error of that module, followed by a 
>> reset of the DAQ.
>> > > >
>> > > >
>> > > > Thus, I am currently unable to test 
>> SIS3316_check_hit_reworked_rebased. Sorry :-(
>> > > >
>> > > >
>> > > >
>> > > >
>> > > > Best greetings
>> > > >
>> > > > Günter
>> > >
>> > >
>> >
>> >
>>
>>
>

More information about the subexp-daq mailing list