Thanks for tracing the problem further down. Now I realized that the CMakeLists.txt for the ROOT analyzer did not contain the usual ROOT flags. I added that and committed the change, so please try again. Here is the diff:

examples/experiment/CMakeLists.txt:

       execute_process(COMMAND root-config --incdir OUTPUT_VARIABLE ROOT_INC)
       execute_process(COMMAND root-config --libs OUTPUT_VARIABLE ROOT_LIBS)
+      execute_process(COMMAND root-config --cflags OUTPUT_VARIABLE ROOT_CFLAGS)
       string(STRIP ${ROOT_LIBS} ROOT_LIBS)
-      string(REGEX REPLACE "\n$" "" ROOT_INC "${ROOT_INC}")
+      string(REGEX REPLACE "\n" "" ROOT_INC ${ROOT_INC})
+      separate_arguments(ROOT_CFLAGS UNIX_COMMAND "${ROOT_CFLAGS}")
       set(CMAKE_CXX_STANDARD 11)

       target_include_directories(analyzer PUBLIC ${ROOT_INC} ${INC_PATH})
+      target_compile_options(analyzer PUBLIC ${ROOT_CFLAGS})
       target_link_libraries(analyzer rmana midas ${ROOT_LIBS} ${LIBS})


Best,
Stefan

> Re: ROOT problem
> 
> I looked into how my ROOT based MIDAS analyzer compiles. It is using the flag -std=c++14.
> MIDAS cmake compiles with -std=gnu++11 with the result:
> 
>  29%] Building CXX object CMakeFiles/rmana.dir/src/mana.cxx.o
> /usr/bin/c++  -D_LARGEFILE64_SOURCE -I/home/pkunz/packages/midas/include -I/home/pkunz/packages/midas/mxml -I/usr/include/root  -O2 -g -DNDEBUG   -DHAVE_ZLIB -DHAVE_FTPLIB -Wall -Wformat=2 -Wno-format-nonliteral -Wno-strict-aliasing -Wuninitialized -Wno-unused-function -DHAVE_ROOT -std=gnu++11 -o CMakeFiles/rmana.dir/src/mana.cxx.o -c /home/pkunz/packages/midas/src/mana.cxx
> In file included from /usr/include/root/TString.h:28,
>                  from /usr/include/root/TCollection.h:29,
>                  from /usr/include/root/TSeqCollection.h:25,
>                  from /usr/include/root/TList.h:25,
>                  from /usr/include/root/TQObject.h:40,
>                  from /usr/include/root/TApplication.h:30,
>                  from /home/pkunz/packages/midas/src/mana.cxx:60:
> /usr/include/root/ROOT/RStringView.hxx:32:37: error: ‘experimental’ in namespace ‘std’ does not name a type
>    32 |    using basic_string_view = ::std::experimental::basic_string_view<_CharT,_Traits>;
> 
> 
> If I change this to
> 
> /usr/bin/c++  -D_LARGEFILE64_SOURCE -I/home/pkunz/packages/midas/include -I/home/pkunz/packages/midas/mxml -I/usr/include/root  -O2 -g -DNDEBUG   -DHAVE_ZLIB -DHAVE_FTPLIB -Wall -Wformat=2 -Wno-format-nonliteral -Wno-strict-aliasing -Wuninitialized -Wno-unused-function -DHAVE_ROOT -std=c++14 -o CMakeFiles/rmana.dir/src/mana.cxx.o -c /home/pkunz/packages/midas/src/mana.cxx
> 
> it compiles without error. I hope this helps. 
> 
> 
> 
> 
> 
> 
> > (please post messages in "plain" mode, they are much easier to answer)
> > 
> > - nvidia problems - this code was contributed by Joseph (I think?), with luck he will look into 
> > this problem.
> > 
> > - ROOT problem - it looks like the error is thrown by the ROOT header files, has nothing to do 
> > with MIDAS?
> > 
> > So what ROOT are you using? I recommend installing ROOT by following instructions at 
> > root.cern.ch.
> > 
> > Perhaps you used the ROOT packages from the EPEL repository? I have seen trouble with 
> > those packages before (miscompiled; important optional features turned off; very old 
> > versions; etc).
> > 
> > P.S.
> > 
> > Historically, ROOT has caused so many reports of "cannot build midas" that I consistently 
> > vote to "remove ROOT support from MIDAS". But Stefan's code for writing MIDAS data into 
> > ROOT files is so neat, cannot throw it away. And some people do use it. So at the latest MIDAS 
> > bash this Summer we decided to keep it.
> > 
> > (Only build targets to use ROOT are the rmlogger executable and the rmana.o object file (and 
> > it's one-man-army library)).
> > 
> > But.
> > 
> > In the past, one could use "make -k" to get past the errors caused by ROOT, everything will 
> > get built and installed, except for the code that failed to build.
> > 
> > Now with cmake, it is "all or nothing", if there is any compilation error, nothing gets installed 
> > into the "bin" directory. So one must discover and use "NO_ROOT=1" (which becomes sticky 
> > until the next "make cclean". Some people are not used to sticky "make" options, I just got 
> > burned by this very thing last week).
> > 
> > Perhaps there is a way to tell cmake to ignore compile errors for rmlogger and rmana.
> > 
> > 
> > K.O.
> > 
> > 
> > 
> > <p>&nbsp;</p>
> > 
> > <table align="center" cellspacing="1" style="border:1px solid #486090; width:98%">
> > 	<tbody>
> > 		<tr>
> > 			<td style="background-color:#486090">Peter Kunz wrote:</td>
> > 		</tr>
> > 		<tr>
> > 			<td style="background-color:#FFFFB0">
> > 			<p>While upgrading to the latest MIDAS version</p>
> > 
> > 			<p>MIDAS version: 2.1 GIT revision: Tue Dec 31 17:40:14 2019 +0100 - midas-
> > 2019-09-i-1-gd93944ce-dirty on branch develop</p>
> > 
> > 			<p>ODB version: 3</p>
> > 
> > 			<p>I encountered two issues using cmake</p>
> > 
> > 			<p>1. on machines with NVIDIA drivers:</p>
> > 
> > 			<div style="background:#eee;border:1px solid #ccc;padding:5px 
> > 10px;">nvml.h: no such file or directory</div>
> > 
> > 			<p>(nvml.h doesn&#39;t seem to be part of the standard nvidia driver 
> > package.)</p>
> > 
> > 			<p>2. Complile including ROOT throws an error with ROOT 6.12/06 on Centos7 
> > and with ROOT 6.18/04 on Fedora 31:</p>
> > 
> > 			<div style="background:#eee;border:1px solid #ccc;padding:5px 10px;">[ 29%] 
> > Building CXX object CMakeFiles/rmana.dir/src/mana.cxx.o In file included from 
> > /usr/include/root/TString.h:28, from /usr/include/root/TCollection.h:29, from 
> > /usr/include/root/TSeqCollection.h:25, from /usr/include/root/TList.h:25, from 
> > /usr/include/root/TQObject.h:40, from /usr/include/root/TApplication.h:30, from 
> > /home/pkunz/packages/midas/src/mana.cxx:60:</div>
> > 
> > 			<div style="background:#eee;border:1px solid #ccc;padding:5px 
> > 10px;">/usr/include/root/ROOT/RStringView.hxx:32:37: error: &lsquo;experimental&rsquo; in 
> > namespace &lsquo;std&rsquo; does not name a type 32 | using basic_string_view = 
> > ::std::experimental::basic_string_view&lt;_CharT,_Traits&gt;;</div>
> > 
> > 			<p>A workaround (which works for me) is to compile with</p>
> > 
> > 			<p><tt>cmake .. -DNO_ROOT=1 -DNO_NVIDIA=1</tt></p>
> > 			</td>
> > 		</tr>
> > 	</tbody>
> > </table>
> > 
> > <p>&nbsp;</p>

> In the past, one could use "make -k" to get past the errors caused by ROOT, everything will 
> get built and installed, except for the code that failed to build.
> 
> Now with cmake, it is "all or nothing", if there is any compilation error, nothing gets installed 
> into the "bin" directory.

That's not correct. Indeed a "make -k" stops after an error, but "make -i" still works. Try it! Event "make install -i" 
installs all executables into the bin directory which successfully compiled even if there is an error.

Stefan

I'm still trying to upgrade my MIDAS system to midas-2019-09-i. Most frontends work fine with the modifications already discussed.
However, I ran into some trouble with the epics frontend. Even with the modifications it throws a lot of warnings and errors (see attached log file). I can reduce the errors with -fpermissive, but the following two errors are persistent:

/home/ays/packages/midas/drivers/device/epics_ca.cxx:167:38: error: ‘ca_create_channel’ was not declared in this scope
            , &(info->caid[i].chan_id))

/home/ays/packages/midas/drivers/device/epics_ca.cxx:178:37: error: ‘ca_create_subscription’ was not declared in this scope
            , &(info->caid[i].evt_id))

This is strange because the functions seem to be declared in base/include/cadef.h along with similar functions that don't throw an error.
I don't know what's going on. The frontend which is almost identical to the example in the midas distribution compiles without warnings or errors under the Midas2017 version. 

> I'm still trying to upgrade my MIDAS system to midas-2019-09-i. Most frontends work fine with the modifications already discussed.
> However, I ran into some trouble with the epics frontend. Even with the modifications it throws a lot of warnings and errors (see attached log file). I can reduce the errors with -fpermissive, but the following two errors are persistent:
> 
> /home/ays/packages/midas/drivers/device/epics_ca.cxx:167:38: error: ‘ca_create_channel’ was not declared in this scope
>             , &(info->caid[i].chan_id))
> 
> /home/ays/packages/midas/drivers/device/epics_ca.cxx:178:37: error: ‘ca_create_subscription’ was not declared in this scope
>             , &(info->caid[i].evt_id))
> 
> This is strange because the functions seem to be declared in base/include/cadef.h along with similar functions that don't throw an error.
> I don't know what's going on. The frontend which is almost identical to the example in the midas distribution compiles without warnings or errors under the Midas2017 version. 

Hi, Peter - it looks like epics_ca.cxx needs to be updated to proper C++ (char* warnings, etc).

As for the "function not declared" errors, it's the C++ way to say "you are calling a function with wrong arguments",
again, something I should be able to fix without too much trouble.

Thank you for reporting this, the midas version of epics_ca.cxx definitely needs fixing.

K.O.

I fixed the compiler errors in epics_ca.cxx, can you try again? (see https://bitbucket.org/tmidas/midas/commits/)

But, I do not see errors with ca_create_channel() and ca_create_subscription().

Maybe I am using the wrong epics. If you still see these errors, let me know what epics you have, I will try the same one.

This is what I do (and I get epics 7.0).

mkdir -p $HOME/git
cd $HOME/git
git clone https://git.launchpad.net/epics-base
cd epics-base
make
ls -l include/cadef.h
ls -l lib/darwin-x86/libca.a # also linux-x86/libca.a

K.O.

> > I'm still trying to upgrade my MIDAS system to midas-2019-09-i. Most frontends work fine with the modifications already discussed.
> > However, I ran into some trouble with the epics frontend. Even with the modifications it throws a lot of warnings and errors (see attached log file). I can reduce the errors with -fpermissive, but the following two errors are persistent:
> > 
> > /home/ays/packages/midas/drivers/device/epics_ca.cxx:167:38: error: ‘ca_create_channel’ was not declared in this scope
> >             , &(info->caid[i].chan_id))
> > 
> > /home/ays/packages/midas/drivers/device/epics_ca.cxx:178:37: error: ‘ca_create_subscription’ was not declared in this scope
> >             , &(info->caid[i].evt_id))
> > 
> > This is strange because the functions seem to be declared in base/include/cadef.h along with similar functions that don't throw an error.
> > I don't know what's going on. The frontend which is almost identical to the example in the midas distribution compiles without warnings or errors under the Midas2017 version. 
> 
> Hi, Peter - it looks like epics_ca.cxx needs to be updated to proper C++ (char* warnings, etc).
> 
> As for the "function not declared" errors, it's the C++ way to say "you are calling a function with wrong arguments",
> again, something I should be able to fix without too much trouble.
> 
> Thank you for reporting this, the midas version of epics_ca.cxx definitely needs fixing.
> 
> K.O.

Hi Konstantin,
I have EPICS Base Release 3.14.8.2 and got your example running with it, though I had to make one change.
For some reason it wouldn't find libca.so, but when I linked the static library libca.a instead, it worked.
Also my midas frontend works now with the updated epics_ca.xx. There was a problem with an old local version of cadef.h
(probably a leftover from some previous testing). After removing it everything compiled well.

Thanks,
Peter


> I fixed the compiler errors in epics_ca.cxx, can you try again? (see https://bitbucket.org/tmidas/midas/commits/)
> 
> But, I do not see errors with ca_create_channel() and ca_create_subscription().
> 
> Maybe I am using the wrong epics. If you still see these errors, let me know what epics you have, I will try the same one.
> 
> This is what I do (and I get epics 7.0).
> 
> mkdir -p $HOME/git
> cd $HOME/git
> git clone https://git.launchpad.net/epics-base
> cd epics-base
> make
> ls -l include/cadef.h
> ls -l lib/darwin-x86/libca.a # also linux-x86/libca.a
> 
> K.O.
> 
> > > I'm still trying to upgrade my MIDAS system to midas-2019-09-i. Most frontends work fine with the modifications already discussed.
> > > However, I ran into some trouble with the epics frontend. Even with the modifications it throws a lot of warnings and errors (see attached log file). I can reduce the errors with -fpermissive, but the following two errors are persistent:
> > > 
> > > /home/ays/packages/midas/drivers/device/epics_ca.cxx:167:38: error: ‘ca_create_channel’ was not declared in this scope
> > >             , &(info->caid[i].chan_id))
> > > 
> > > /home/ays/packages/midas/drivers/device/epics_ca.cxx:178:37: error: ‘ca_create_subscription’ was not declared in this scope
> > >             , &(info->caid[i].evt_id))
> > > 
> > > This is strange because the functions seem to be declared in base/include/cadef.h along with similar functions that don't throw an error.
> > > I don't know what's going on. The frontend which is almost identical to the example in the midas distribution compiles without warnings or errors under the Midas2017 version. 
> > 
> > Hi, Peter - it looks like epics_ca.cxx needs to be updated to proper C++ (char* warnings, etc).
> > 
> > As for the "function not declared" errors, it's the C++ way to say "you are calling a function with wrong arguments",
> > again, something I should be able to fix without too much trouble.
> > 
> > Thank you for reporting this, the midas version of epics_ca.cxx definitely needs fixing.
> > 
> > K.O.

> I have EPICS Base Release 3.14.8.2 and got your example running with it...

Ok, good.

> I had to make one change.
> For some reason it wouldn't find libca.so, but when I linked the static library libca.a instead, it worked.

Yes, libca.a is what I see in my epics, too. I recommend static linking (libxxx.a instead of libxxx.so) as it eliminates
all possible problems with "cannot find the shared library" and "found the wrong shared library" and "found
a shared library from a different version of epics".

> Also my midas frontend works now with the updated epics_ca.xx. There was a problem with an old local version of cadef.h
> (probably a leftover from some previous testing). After removing it everything compiled well.

Ok. good. I also see some old epics header files in the midas repository. I guess I should remove them.

K.O.


> 
> Thanks,
> Peter
> 
> 
> > I fixed the compiler errors in epics_ca.cxx, can you try again? (see https://bitbucket.org/tmidas/midas/commits/)
> > 
> > But, I do not see errors with ca_create_channel() and ca_create_subscription().
> > 
> > Maybe I am using the wrong epics. If you still see these errors, let me know what epics you have, I will try the same one.
> > 
> > This is what I do (and I get epics 7.0).
> > 
> > mkdir -p $HOME/git
> > cd $HOME/git
> > git clone https://git.launchpad.net/epics-base
> > cd epics-base
> > make
> > ls -l include/cadef.h
> > ls -l lib/darwin-x86/libca.a # also linux-x86/libca.a
> > 
> > K.O.
> > 
> > > > I'm still trying to upgrade my MIDAS system to midas-2019-09-i. Most frontends work fine with the modifications already discussed.
> > > > However, I ran into some trouble with the epics frontend. Even with the modifications it throws a lot of warnings and errors (see attached log file). I can reduce the errors with -fpermissive, but the following two errors are persistent:
> > > > 
> > > > /home/ays/packages/midas/drivers/device/epics_ca.cxx:167:38: error: ‘ca_create_channel’ was not declared in this scope
> > > >             , &(info->caid[i].chan_id))
> > > > 
> > > > /home/ays/packages/midas/drivers/device/epics_ca.cxx:178:37: error: ‘ca_create_subscription’ was not declared in this scope
> > > >             , &(info->caid[i].evt_id))
> > > > 
> > > > This is strange because the functions seem to be declared in base/include/cadef.h along with similar functions that don't throw an error.
> > > > I don't know what's going on. The frontend which is almost identical to the example in the midas distribution compiles without warnings or errors under the Midas2017 version. 
> > > 
> > > Hi, Peter - it looks like epics_ca.cxx needs to be updated to proper C++ (char* warnings, etc).
> > > 
> > > As for the "function not declared" errors, it's the C++ way to say "you are calling a function with wrong arguments",
> > > again, something I should be able to fix without too much trouble.
> > > 
> > > Thank you for reporting this, the midas version of epics_ca.cxx definitely needs fixing.
> > > 
> > > K.O.

>
> > I have EPICS Base Release 3.14.8.2 ...
>

For the record, according to https://epics.anl.gov/

the 3.14 branch is "Closed, please upgrade!"
the 3.15 branch seems to be the current "stable" branch, and
the 7.0 branch is the "current release"

I forget which version of EPICS is currently recommended by the TRIUMF controls group. I will have to go and ask them.

K.O.

> Complile including ROOT throws an error with ROOT 6.18/04 on Fedora 31 ...
> /usr/include/root/ROOT/RStringView.hxx:32:37: error: ‘experimental’ in namespace ‘std’ ...

I was puzzled by this error about "experimental" in ROOT header files.

All ROOT kits that I have are configured and built with c++11, and I do not see this error. I now guess that you
only saw this on Fedora 31 with gcc 9 and ROOT that was configured for c++14 or c++17.

We now realize that MIDAS cmake does not use the ROOT CFLAGS when building code that uses ROOT,
and this is probably a mistake. Stefan added a fix to use ROOT CFLAGS to build the analyzer,
but I am not sure if this is sufficient, rmana and rmlogger probably also should use the ROOT CFLAGS.

In any case, we are setup for an eventual collision between MIDAS CFLAGS (-std=gnu++11) and ROOT CFLAGS (-std=something else).

K.O.

> In any case, we are setup for an eventual collision between MIDAS CFLAGS (-std=gnu++11) and ROOT CFLAGS (-std=something else).

Are there good reasons to not compile MIDAS with set(CMAKE_CXX_STANDARD 14)? So far this was an easier "fix" for me than to recompile ROOT with c++11.

> > In any case, we are setup for an eventual collision between MIDAS CFLAGS (-std=gnu++11) and ROOT CFLAGS (-std=something else).
> 
> Are there good reasons to not compile MIDAS with set(CMAKE_CXX_STANDARD 14)? So far this was an easier "fix" for me than to recompile ROOT with c++11.

Yes. We agreed to base MIDAS on C++11, not C++14, since C++11 is available on all relevant systems. By putting that requirement explicitly into the CMake file, we know that it 
correctly compiles with C++11. Of course feel free to overwrite this in your local installation.

Stefan

> > In any case, we are setup for an eventual collision between MIDAS CFLAGS (-std=gnu++11) and ROOT CFLAGS (-std=something else).
> 
> Are there good reasons to not compile MIDAS with set(CMAKE_CXX_STANDARD 14)? So far this was an easier "fix" for me than to recompile ROOT with c++11.

We already have trouble enough even with requiring C++11:

We still run a number of SL6 DAQ stations at TRIUMF, and it is well known that SL6 gcc does not even do c++11.

The suggested solution to "simply just use the devtoolset-8" compilers turned out to be missing a critical 32-bit
library and cannot build midas for 32-bit VME processors. (straight or cross-compiled).

(A solution for this can be kludged using unofficial packages: https://midas.triumf.ca/elog/Midas/1748)

No, we cannot "just upgrade" SL6 and no, we cannot "just replace" the 32-bit-only VME processors.

If we move the midas requirements from c++11 to c++14, again, we have to see what compilers
are available on what platforms, and I am sure again we will miss some important case.

And then, where does this train stop?

I see c++17 has some really nice to use improvements. And I see c++20 added even some more nice improvements.

I say we follow ROOT: https://root.cern.ch/supported-platforms

> Supported Platforms
> ROOT 6
> The compiler needs to support -std=c++11 to be able to build version 6.

K.O.

> > In any case, we are setup for an eventual collision between MIDAS CFLAGS (-std=gnu++11) and ROOT CFLAGS (-std=something else).
> 
> Are there good reasons to not compile MIDAS with set(CMAKE_CXX_STANDARD 14)? So far this was an easier "fix" for me than to recompile ROOT with c++11.

Per  https://root.cern.ch/supported-platforms, ROOTv6 is C++11.

Binary kits distributed from root.cern.ch seem to be built with c++11.

Where does ROOT built with c++14 come from?

K.O.

Dear all,

we are trying to build Midas events on FPGAs and send them directly to the midas
ring buffer via copy_n. According to the wiki
https://midas.triumf.ca/MidasWiki/index.php/Event_Structure Event Data Size is:
"The event data size contains the size of the event in bytes excluding the
header." and All Bank Size is: "Size in bytes of the following data plus the
size of the bank header". So are they actually the same or what header is the
header in the first sentence also the bank header?

Cheers,
Marius

 

Thanks for pointing out this error. The "All Bank Size" contains the size of all banks including their 
bank headers, but NOT the global bank header itself. I modified the documentation accordingly.

If you want to study the C code which tells you how to fill these headers, look at midas.cxx line 
14788.

Best,
Stefan

> Dear all,
> 
> we are trying to build Midas events on FPGAs and send them directly to the midas
> ring buffer via copy_n. According to the wiki
> https://midas.triumf.ca/MidasWiki/index.php/Event_Structure Event Data Size is:
> "The event data size contains the size of the event in bytes excluding the
> header." and All Bank Size is: "Size in bytes of the following data plus the
> size of the bank header". So are they actually the same or what header is the
> header in the first sentence also the bank header?
> 
> Cheers,
> Marius
> 
>  

Dear all,

we creating Midas events directly inside a FPGA and send them off via DMA into the PC RAM. For reading out this RAM via Midas the FPGA sends as a pointer where it has written the last 4kB of data. We use this pointer for telling the ring buffer of midas where the new events are. The buffer looks something like:

// event 1
dma_buf[0] = 0x00000001; // Trigger and Event ID
dma_buf[1] = 0x00000001; // Serial number
dma_buf[2] = TIME; // time
dma_buf[3] = 18*4-4*4; // event size
dma_buf[4] = 18*4-6*4; // all bank size
dma_buf[5] = 0x11; // flags
// bank 0
dma_buf[6] = 0x46454230; // bank name
dma_buf[7] = 0x6; // bank type TID_DWORD
dma_buf[8] = 0x3*4; // data size
dma_buf[9] = 0xAFFEAFFE; // data
dma_buf[10] = 0xAFFEAFFE; // data
dma_buf[11] = 0xAFFEAFFE; // data
// bank 1
dma_buf[12] = 0x1; // bank name
dma_buf[12] = 0x46454231; // bank name
dma_buf[13] = 0x6; // bank type TID_DWORD
dma_buf[14] = 0x3*4; // data size
dma_buf[15] = 0xAFFEAFFE; // data
dma_buf[16] = 0xAFFEAFFE; // data
dma_buf[17] = 0xAFFEAFFE; // data

// event 2
.....

dma_buf[fpga_pointer] = 0xXXXXXXXX;


And we do something like:

while{true}
   // obtain buffer space
   status = rb_get_wp(rbh, (void **)&pdata, 10);
   fpga_pointer = fpga.read_last_data_add();

   wlen = last_fpga_pointer - fpga_pointer; \\ in 32 bit words
   copy_n(&dma_buf[last_fpga_pointer], wlen, pdata);
   rb_status = rb_increment_wp(rbh, wlen * 4); \\ in byte

   last_fpga_pointer = fpga_pointer;

Leaving the case out where the dma_buf wrap around this works fine for a small data rate. But if we increase the rate the fpga_pointer also increases really fast and wlen gets quite big. Actually it gets bigger then max_event_size which is checked in rb_increment_wp leading to an error. 

The problem now is that the event size is actually not to big but since we have multi events in the buffer which are read by midas in one step. So we think in this case the function rb_increment_wp is comparing actually the wrong thing. Also increasing the max_event_size does not help.

Remark: dma_buf is volatile so memcpy is not possible here.

Cheers,
Marius

The rb_xxx functions are (thoroughly tested) robust against high data rate given that you use them as intended:

1) Once you create the ring buffer via rb_create(), specify the maximum event size (midas event, not bank size!). Later there is not protection any more, so if you obtain pdata from rb_get_wp, you can of course write 4GB to pdata, overwriting everything in your memory, causing a fatal crash. It's your duty not to write more bytes into pdata then what you specified in rb_create()

2) Once you obtain a write pointer to the ring buffer via rb_get_wp(), this function might fail when the receiving side reads data slower than the producing side, simply because the buffer is full. In that case the producing side was to wait to get new buffer space. If you call to rb_get_wp() returns DB_TIMEOUT, it means that the function did not obtain enough free space for the next event. In that case you have to wait (like ss_sleep(10)) and try again. Only when rb_get_wp() returns DB_SUCCESS, you are allowed to write into pdata (up to the maximum event size specified in rb_create() of course).

/Stefan


> Dear all,
> 
> we creating Midas events directly inside a FPGA and send them off via DMA into the PC RAM. For reading out this RAM via Midas the FPGA sends as a pointer where it has written the last 4kB of data. We use this pointer for telling the ring buffer of midas where the new events are. The buffer looks something like:
> 
> // event 1
> dma_buf[0] = 0x00000001; // Trigger and Event ID
> dma_buf[1] = 0x00000001; // Serial number
> dma_buf[2] = TIME; // time
> dma_buf[3] = 18*4-4*4; // event size
> dma_buf[4] = 18*4-6*4; // all bank size
> dma_buf[5] = 0x11; // flags
> // bank 0
> dma_buf[6] = 0x46454230; // bank name
> dma_buf[7] = 0x6; // bank type TID_DWORD
> dma_buf[8] = 0x3*4; // data size
> dma_buf[9] = 0xAFFEAFFE; // data
> dma_buf[10] = 0xAFFEAFFE; // data
> dma_buf[11] = 0xAFFEAFFE; // data
> // bank 1
> dma_buf[12] = 0x1; // bank name
> dma_buf[12] = 0x46454231; // bank name
> dma_buf[13] = 0x6; // bank type TID_DWORD
> dma_buf[14] = 0x3*4; // data size
> dma_buf[15] = 0xAFFEAFFE; // data
> dma_buf[16] = 0xAFFEAFFE; // data
> dma_buf[17] = 0xAFFEAFFE; // data
> 
> // event 2
> .....
> 
> dma_buf[fpga_pointer] = 0xXXXXXXXX;
> 
> 
> And we do something like:
> 
> while{true}
>    // obtain buffer space
>    status = rb_get_wp(rbh, (void **)&pdata, 10);
>    fpga_pointer = fpga.read_last_data_add();
> 
>    wlen = last_fpga_pointer - fpga_pointer; \\ in 32 bit words
>    copy_n(&dma_buf[last_fpga_pointer], wlen, pdata);
>    rb_status = rb_increment_wp(rbh, wlen * 4); \\ in byte
> 
>    last_fpga_pointer = fpga_pointer;
> 
> Leaving the case out where the dma_buf wrap around this works fine for a small data rate. But if we increase the rate the fpga_pointer also increases really fast and wlen gets quite big. Actually it gets bigger then max_event_size which is checked in rb_increment_wp leading to an error. 
> 
> The problem now is that the event size is actually not to big but since we have multi events in the buffer which are read by midas in one step. So we think in this case the function rb_increment_wp is comparing actually the wrong thing. Also increasing the max_event_size does not help.
> 
> Remark: dma_buf is volatile so memcpy is not possible here.
> 
> Cheers,
> Marius

The rb_xxx function are (thoroughly tested!) robust against high data rate given that you use them as intended:

1) Once you create the ring buffer via rb_create(), specify the maximum event size (overall event size, not bank size!). Later there is no protection any more, so if you obtain pdata from rb_get_wp, you can of course write 4GB to pdata, overwriting everything in your memory, causing a total crash. It's your responsibility to not write more bytes into pdata then 
what you specified as max event size in rb_create()

2) Once you obtain a write pointer to the ring buffer via rb_get_wp, this function might fail when the receiving side reads data slower than the producing side, simply because the buffer is full. In that case the producing side has to wait until space is freed up in the buffer by the receiving side. If your call to rb_get_wp returns DB_TIMEOUT, it means that the 
function did not obtain enough free space for the next event. In that case you have to wait (like ss_sleep(10)) and try again, until you succeed. Only when rb_get_wp() returns DB_SUCCESS, you are allowed to write into pdata, up to the maximum event size specified in rb_create of course. I don't see this behaviour in your code. You would need something 
like

do {
   status = rb_get_wp(rbh, (void **)&pdata, 10);
   if (status == DB_TIMEOUT)
      ss_sleep(10);
   } while (status == DB_TIMEOUT);

Best,
Stefan


> Dear all,
> 
> we creating Midas events directly inside a FPGA and send them off via DMA into the PC RAM. For reading out this RAM via Midas the FPGA sends as a pointer where it has written the last 4kB of data. We use this pointer for telling the ring buffer of midas where the new events are. The buffer looks something like:
> 
> // event 1
> dma_buf[0] = 0x00000001; // Trigger and Event ID
> dma_buf[1] = 0x00000001; // Serial number
> dma_buf[2] = TIME; // time
> dma_buf[3] = 18*4-4*4; // event size
> dma_buf[4] = 18*4-6*4; // all bank size
> dma_buf[5] = 0x11; // flags
> // bank 0
> dma_buf[6] = 0x46454230; // bank name
> dma_buf[7] = 0x6; // bank type TID_DWORD
> dma_buf[8] = 0x3*4; // data size
> dma_buf[9] = 0xAFFEAFFE; // data
> dma_buf[10] = 0xAFFEAFFE; // data
> dma_buf[11] = 0xAFFEAFFE; // data
> // bank 1
> dma_buf[12] = 0x1; // bank name
> dma_buf[12] = 0x46454231; // bank name
> dma_buf[13] = 0x6; // bank type TID_DWORD
> dma_buf[14] = 0x3*4; // data size
> dma_buf[15] = 0xAFFEAFFE; // data
> dma_buf[16] = 0xAFFEAFFE; // data
> dma_buf[17] = 0xAFFEAFFE; // data
> 
> // event 2
> .....
> 
> dma_buf[fpga_pointer] = 0xXXXXXXXX;
> 
> 
> And we do something like:
> 
> while{true}
>    // obtain buffer space
>    status = rb_get_wp(rbh, (void **)&pdata, 10);
>    fpga_pointer = fpga.read_last_data_add();
> 
>    wlen = last_fpga_pointer - fpga_pointer; \\ in 32 bit words
>    copy_n(&dma_buf[last_fpga_pointer], wlen, pdata);
>    rb_status = rb_increment_wp(rbh, wlen * 4); \\ in byte
> 
>    last_fpga_pointer = fpga_pointer;
> 
> Leaving the case out where the dma_buf wrap around this works fine for a small data rate. But if we increase the rate the fpga_pointer also increases really fast and wlen gets quite big. Actually it gets bigger then max_event_size which is checked in rb_increment_wp leading to an error. 
> 
> The problem now is that the event size is actually not to big but since we have multi events in the buffer which are read by midas in one step. So we think in this case the function rb_increment_wp is comparing actually the wrong thing. Also increasing the max_event_size does not help.
> 
> Remark: dma_buf is volatile so memcpy is not possible here.
> 
> Cheers,
> Marius

Hello Stefan,
is there a difference for the later data processing (after writing the ring buffer blocks)
if we write single events or multiple in one rb_get_wp - memcopy - rb_increment_wp cycle?
Both Marius and me have seen some inconsistencies in the number of events produced that is reported in the status page when writing multiple events in one go,
so I was wondering if this is due to us treating the buffer badly or the way midas handles the events after that.

Given that we produce the full event in our (FPGA) domain, an option would be to always copy one event from the dma to the midas-system buffer in a loop.
The question is if there is a difference (for midas) between
[pseudo code, much simplified]

while(dma_read_index < last_dma_write_index){
  if(rb_get_wp(pdata)!=SUCCESS){
    dma_read_index+=event_size;
    continue;   
  }
  copy_n(dma_buffer, pdata, event_size);
  rb_increment_wp(event_size);
  dma_read_index+=event_size;
} 

and

while(dma_read_index < last_dma_write_index){
  if(rb_get_wp(pdata)!=SUCCESS){
     ...
  };
  total_size=max_n_events_that_fit_in_rb_block();
  copy_n(dma_buffer, pdata, total_size);
  rb_increment_wp(total_size);
  dma_read_index+=total_size;
}

Cheers,
Konrad

> The rb_xxx function are (thoroughly tested!) robust against high data rate given that you use them as intended:
> 
> 1) Once you create the ring buffer via rb_create(), specify the maximum event size (overall event size, not bank size!). Later there is no protection any more, so if you obtain pdata from rb_get_wp, you can of course write 4GB to pdata, overwriting everything in your memory, causing a total crash. It's your responsibility to not write more bytes into pdata then 
> what you specified as max event size in rb_create()
> 
> 2) Once you obtain a write pointer to the ring buffer via rb_get_wp, this function might fail when the receiving side reads data slower than the producing side, simply because the buffer is full. In that case the producing side has to wait until space is freed up in the buffer by the receiving side. If your call to rb_get_wp returns DB_TIMEOUT, it means that the 
> function did not obtain enough free space for the next event. In that case you have to wait (like ss_sleep(10)) and try again, until you succeed. Only when rb_get_wp() returns DB_SUCCESS, you are allowed to write into pdata, up to the maximum event size specified in rb_create of course. I don't see this behaviour in your code. You would need something 
> like
> 
> do {
>    status = rb_get_wp(rbh, (void **)&pdata, 10);
>    if (status == DB_TIMEOUT)
>       ss_sleep(10);
>    } while (status == DB_TIMEOUT);
> 
> Best,
> Stefan
> 
> 
> > Dear all,
> > 
> > we creating Midas events directly inside a FPGA and send them off via DMA into the PC RAM. For reading out this RAM via Midas the FPGA sends as a pointer where it has written the last 4kB of data. We use this pointer for telling the ring buffer of midas where the new events are. The buffer looks something like:
> > 
> > // event 1
> > dma_buf[0] = 0x00000001; // Trigger and Event ID
> > dma_buf[1] = 0x00000001; // Serial number
> > dma_buf[2] = TIME; // time
> > dma_buf[3] = 18*4-4*4; // event size
> > dma_buf[4] = 18*4-6*4; // all bank size
> > dma_buf[5] = 0x11; // flags
> > // bank 0
> > dma_buf[6] = 0x46454230; // bank name
> > dma_buf[7] = 0x6; // bank type TID_DWORD
> > dma_buf[8] = 0x3*4; // data size
> > dma_buf[9] = 0xAFFEAFFE; // data
> > dma_buf[10] = 0xAFFEAFFE; // data
> > dma_buf[11] = 0xAFFEAFFE; // data
> > // bank 1
> > dma_buf[12] = 0x1; // bank name
> > dma_buf[12] = 0x46454231; // bank name
> > dma_buf[13] = 0x6; // bank type TID_DWORD
> > dma_buf[14] = 0x3*4; // data size
> > dma_buf[15] = 0xAFFEAFFE; // data
> > dma_buf[16] = 0xAFFEAFFE; // data
> > dma_buf[17] = 0xAFFEAFFE; // data
> > 
> > // event 2
> > .....
> > 
> > dma_buf[fpga_pointer] = 0xXXXXXXXX;
> > 
> > 
> > And we do something like:
> > 
> > while{true}
> >    // obtain buffer space
> >    status = rb_get_wp(rbh, (void **)&pdata, 10);
> >    fpga_pointer = fpga.read_last_data_add();
> > 
> >    wlen = last_fpga_pointer - fpga_pointer; \\ in 32 bit words
> >    copy_n(&dma_buf[last_fpga_pointer], wlen, pdata);
> >    rb_status = rb_increment_wp(rbh, wlen * 4); \\ in byte
> > 
> >    last_fpga_pointer = fpga_pointer;
> > 
> > Leaving the case out where the dma_buf wrap around this works fine for a small data rate. But if we increase the rate the fpga_pointer also increases really fast and wlen gets quite big. Actually it gets bigger then max_event_size which is checked in rb_increment_wp leading to an error. 
> > 
> > The problem now is that the event size is actually not to big but since we have multi events in the buffer which are read by midas in one step. So we think in this case the function rb_increment_wp is comparing actually the wrong thing. Also increasing the max_event_size does not help.
> > 
> > Remark: dma_buf is volatile so memcpy is not possible here.
> > 
> > Cheers,
> > Marius

rb_xxx functions are midas event agnostic. The receiving side in mfe.cxx (lines 1418 in receive_trigger_event) however pulls one event at a time. If you 
have some inconsistency I would put some debugging code there.

Stefan

> Hello Stefan,
> is there a difference for the later data processing (after writing the ring buffer blocks)
> if we write single events or multiple in one rb_get_wp - memcopy - rb_increment_wp cycle?
> Both Marius and me have seen some inconsistencies in the number of events produced that is reported in the status page when writing multiple 
events in one go,
> so I was wondering if this is due to us treating the buffer badly or the way midas handles the events after that.
> 
> Given that we produce the full event in our (FPGA) domain, an option would be to always copy one event from the dma to the midas-system buffer 
in a loop.
> The question is if there is a difference (for midas) between
> [pseudo code, much simplified]
> 
> while(dma_read_index < last_dma_write_index){
>   if(rb_get_wp(pdata)!=SUCCESS){
>     dma_read_index+=event_size;
>     continue;   
>   }
>   copy_n(dma_buffer, pdata, event_size);
>   rb_increment_wp(event_size);
>   dma_read_index+=event_size;
> } 
> 
> and
> 
> while(dma_read_index < last_dma_write_index){
>   if(rb_get_wp(pdata)!=SUCCESS){
>      ...
>   };
>   total_size=max_n_events_that_fit_in_rb_block();
>   copy_n(dma_buffer, pdata, total_size);
>   rb_increment_wp(total_size);
>   dma_read_index+=total_size;
> }
> 
> Cheers,
> Konrad
> 
> > The rb_xxx function are (thoroughly tested!) robust against high data rate given that you use them as intended:
> > 
> > 1) Once you create the ring buffer via rb_create(), specify the maximum event size (overall event size, not bank size!). Later there is no protection 
any more, so if you obtain pdata from rb_get_wp, you can of course write 4GB to pdata, overwriting everything in your memory, causing a total crash. 
It's your responsibility to not write more bytes into pdata then 
> > what you specified as max event size in rb_create()
> > 
> > 2) Once you obtain a write pointer to the ring buffer via rb_get_wp, this function might fail when the receiving side reads data slower than the 
producing side, simply because the buffer is full. In that case the producing side has to wait until space is freed up in the buffer by the receiving side. 
If your call to rb_get_wp returns DB_TIMEOUT, it means that the 
> > function did not obtain enough free space for the next event. In that case you have to wait (like ss_sleep(10)) and try again, until you succeed. 
Only when rb_get_wp() returns DB_SUCCESS, you are allowed to write into pdata, up to the maximum event size specified in rb_create of course. I 
don't see this behaviour in your code. You would need something 
> > like
> > 
> > do {
> >    status = rb_get_wp(rbh, (void **)&pdata, 10);
> >    if (status == DB_TIMEOUT)
> >       ss_sleep(10);
> >    } while (status == DB_TIMEOUT);
> > 
> > Best,
> > Stefan
> > 
> > 
> > > Dear all,
> > > 
> > > we creating Midas events directly inside a FPGA and send them off via DMA into the PC RAM. For reading out this RAM via Midas the FPGA 
sends as a pointer where it has written the last 4kB of data. We use this pointer for telling the ring buffer of midas where the new events are. The 
buffer looks something like:
> > > 
> > > // event 1
> > > dma_buf[0] = 0x00000001; // Trigger and Event ID
> > > dma_buf[1] = 0x00000001; // Serial number
> > > dma_buf[2] = TIME; // time
> > > dma_buf[3] = 18*4-4*4; // event size
> > > dma_buf[4] = 18*4-6*4; // all bank size
> > > dma_buf[5] = 0x11; // flags
> > > // bank 0
> > > dma_buf[6] = 0x46454230; // bank name
> > > dma_buf[7] = 0x6; // bank type TID_DWORD
> > > dma_buf[8] = 0x3*4; // data size
> > > dma_buf[9] = 0xAFFEAFFE; // data
> > > dma_buf[10] = 0xAFFEAFFE; // data
> > > dma_buf[11] = 0xAFFEAFFE; // data
> > > // bank 1
> > > dma_buf[12] = 0x1; // bank name
> > > dma_buf[12] = 0x46454231; // bank name
> > > dma_buf[13] = 0x6; // bank type TID_DWORD
> > > dma_buf[14] = 0x3*4; // data size
> > > dma_buf[15] = 0xAFFEAFFE; // data
> > > dma_buf[16] = 0xAFFEAFFE; // data
> > > dma_buf[17] = 0xAFFEAFFE; // data
> > > 
> > > // event 2
> > > .....
> > > 
> > > dma_buf[fpga_pointer] = 0xXXXXXXXX;
> > > 
> > > 
> > > And we do something like:
> > > 
> > > while{true}
> > >    // obtain buffer space
> > >    status = rb_get_wp(rbh, (void **)&pdata, 10);
> > >    fpga_pointer = fpga.read_last_data_add();
> > > 
> > >    wlen = last_fpga_pointer - fpga_pointer; \\ in 32 bit words
> > >    copy_n(&dma_buf[last_fpga_pointer], wlen, pdata);
> > >    rb_status = rb_increment_wp(rbh, wlen * 4); \\ in byte
> > > 
> > >    last_fpga_pointer = fpga_pointer;
> > > 
> > > Leaving the case out where the dma_buf wrap around this works fine for a small data rate. But if we increase the rate the fpga_pointer also 
increases really fast and wlen gets quite big. Actually it gets bigger then max_event_size which is checked in rb_increment_wp leading to an error. 
> > > 
> > > The problem now is that the event size is actually not to big but since we have multi events in the buffer which are read by midas in one step. 
So we think in this case the function rb_increment_wp is comparing actually the wrong thing. Also increasing the max_event_size does not help.
> > > 
> > > Remark: dma_buf is volatile so memcpy is not possible here.
> > > 
> > > Cheers,
> > > Marius