Jump to content

Philipp A Hartmann

Members
  • Content Count

    490
  • Joined

  • Last visited

  • Days Won

    118

Philipp A Hartmann last won the day on May 8

Philipp A Hartmann had the most liked content!

About Philipp A Hartmann

  • Rank
    Advanced Member

Profile Information

  • Gender
    Not Telling
  • Location
    Duisburg, DE

Recent Profile Visitors

2,144 profile views
  1. I gave it a shot locally and noticed that there are indeed quite some compiler warnings due to differences between __declspec(...) and __attribute__((visibility(...))). Most of these warnings can be ignored for now, but are certainly annoying. Additionally, you need to explicitly mark the following with SC_API as well: // src/sysc/communication/sc_port.h template <class IF> class SC_API sc_port_b template <class IF> class SC_API sc_port // src/sysc/communication/sc_prim_channel.h class SC_API sc_prim_channel_registry // src/sysc/communication/sc_signal.h template< class T, sc_writer_policy POL > class SC_API sc_signal_t template< class T, sc_writer_policy POL > class SC_API sc_signal // src/sysc/datatypes/sc_proxy.h template <class X> class SC_API sc_proxy // src/sysc/datatypes/sc_proxy.h extern "C" SC_API int sc_main(int argc, char* argv[]) // src/sysc/utils/sc_temporary.h template<class T> class SC_API sc_vpool For compiling SystemC itself with -fvisibility=hidden, some further changes to the .cpp files are needed. And some of the above changes might break the Windows DLL support. But for me, the above changes work with the examples in the SystemC distribution (i.e. "gmake check CXXFLAGS='-fvisibility=hidden"). Long story short: The feature to SystemC build models with limited symbol visibility will require more changes and testing. I will forward this topic to the Language Working Group for further discussion.
  2. I'm not suggesting to export all of your classes, but the SystemC class (templates) used in your code. By reusing the macro annotations for Windows, you can try changing src/sysc/kernel/sc_cmnhdr.h and redefine SC_API for your platform # define SC_API __attribute__((visibility("default"))) (with this, you can even try building SystemC itself with -fvisibility=hidden).
  3. This is likely caused by the "hidden" RTTI information, messing with dynamic_cast and typeid usage in SystemC, see e.g. https://gcc.gnu.org/wiki/Visibility. You at least need to export all SystemC (template) classes from your shared library. Greetings from Duisburg, Philipp
  4. Philipp A Hartmann

    send array over tlm 2.0 fifio

    These classes are templates themselves. You need to pass the type of the values as parameter, for example tlm_fifo< std::array<int,4> > // a FIFO sending 4-element arrays of integers /Philipp
  5. Philipp A Hartmann

    send array over tlm 2.0 fifio

    Use an std::vector or std::array as value type in your FIFO instead of a C-style array.
  6. Philipp A Hartmann

    Build error SystemC 2.3.3 with C++17

    The INSTALL notes currently say the following about the SC_CPLUSPLUS preprocessor switch (emphasis mine): This means, that you can explicitly opt-out of some SystemC features, that would generally be supported by your compiler's C++ language support. Selecting a newer C++ standard version than supported by your compiler (setup) will typically not work - as you have experienced. That said, adding a preprocessor flag on the compiler command-line does not affect the language mode of the compiler. How to switch your compiler to C++17 mode is outside of the scope of SystemC. Your compiler (version) seems to require the -std=c++17 switch to enable this mode. The SC_CPLUSPLUS flag is then not needed as it will be set automatically to the selected version. Hope that helps, Philipp
  7. Philipp A Hartmann

    runtime error :- Fatal: (F4) assertion failed: iface != 0

    The issue is likely caused because you access a port (via -> or for example calling functions like .read()) already inside the module constructor. You should only access ports after binding has completed, this means from within a SystemC process or in end_of_elaboration() / start_of_simulation() callbacks. Hope that helps, Philipp
  8. From my understanding, the "before_end_of_elaboration" limitation got fixed in SystemC 2.3.1 (and a regression introduced by this was fixed in 2.3.2/2.3.3). /Philipp
  9. Yes, TLM sockets are dependent on SystemC and can only be used inside of a SystemC module. This is because they are implemented via SystemC ports/exports, which can itself only be instantiated in SystemC modules as per the SystemC standard. If you want to bridge into a pure C++ model, you can pass the corresponding interface pointers into your C++ model via your own code (via get_interface() after elaboration). This way, you only depend on sc_interface from SystemC (which is a base class of all SystemC/TLM interface classes) inside your C++ model. I'm not sure, what you are trying to achieve, though. Why can't you use SystemC (which is just C++ at the end anyway)? Greetings from Duisburg, Philipp
  10. Philipp A Hartmann

    how can i know which event triggered a process?

    SystemC 2.3.2 (and later) does support a query whether an event was triggered in the previous delta cycle (or immediately in the current evaluation cycle). This is similar to the sc_signal<>::event() function. With this, you can ask every element in your array, whether it was triggered and may have caused a wakeup of the process. Of course, multiple events can have triggered simultaneously and will only cause a single wakeup of the process. for( const auto ev& : e ) { if ( ev.triggered() ) { // ... } } Hope that helps, Philipp
  11. Philipp A Hartmann

    Killing a process with an included sc_event_[and|or]_list

    Thanks! I can reproduce the behavior and verified that removing the dynamic sensitivity in sc_thread_process::kill_process fixes the issue: void sc_thread_process::kill_process(sc_descendant_inclusion_info descendants ) { // ... if ( sc_is_running() && m_has_stack ) { m_throw_status = THROW_KILL; m_wait_cycle_n = 0; remove_dynamic_events(); // <-- add this line to avoid the exception simcontext()->preempt_with(this); } // ... } I'm not sure, if it is necessary to do the same for the static sensitivity. At least I haven't come up with a similar scenario, where the error is actually "incorrect".
  12. Philipp A Hartmann

    Killing a process with an included sc_event_[and|or]_list

    I did write the above comment and the case where a (e.g. temporary) list lives inside a thread is certainly one of the reasons why we need a check for premature destruction. Due to the nature how event lists are internally handled, we will run into a memory corruption if list is destroyed while a process is still waiting for it. What was not considered is the explicit kill() of a process that is current waiting on such a list. The easiest fix would be to cancel (dynamic) sensitivity before doing the stack unwinding when killing the process. @Stephan Gerth: Can you submit a pull-request with your example above as testcase against the SystemC regressions repository?
  13. Philipp A Hartmann

    wait() is not allowed inside run_phase

    If you use SC_JOIN or your SC_JOIN_ANY, you want to keep the run_phase active until all (or at least one) of the threads terminate(s). Why can't you raise an objection here? You will drop the objection after the join, so the phase can successfully complete. (Without joining, it's fine not to raise an objection).
  14. Philipp A Hartmann

    wait() is not allowed inside run_phase

    The error looks strange to me, as terminated_events is not destroyed before exiting run_phase, i.e. after the wait() has returned. However, I don't know how the objections are implemented in UVM SystemC and whether or how wait() is supported without raising an objection. Can you check with a debugger, if an exception is thrown while your run_phase function is inside the wait()? Hope that helps, Philipp
  15. Philipp A Hartmann

    wait() is not allowed inside run_phase

    Can you show the output of your simulation? And maybe add sc_get_current_process_handle().name() to the log statements to include the process name in the output. I would expect to at last see the output after the wait() in fun2. Secondly, you should check for (*it).terminated() and only kill() running processes in the termination loop.
×