> On June 23, 2016, 7:09 p.m., Matthias Jung wrote: > > Dear Christian, > > > > thank you for this nice contribution. I will have a look on your patch and > > maybe I can help you fixing your issue. > > > > Regards > > Matthias > > Andreas Hansson wrote: > Christian, Matthias, is there any sensible way we could actually get some > of this functionality tested? Including SystemC in ext seems problematic due > to the non-BSD compatible license, but perhaps we should do something like we > do with protobuf, and check if the host HAS_SYSTEMC or similar, and if so > compile the extra bits, and run the extra tests etc. > > This is tremendously useful functionality, and it feels like it is bolted > on rather than properly integrated at the moment (not because of the > functionality, but due to the non-default use). Any thoughts or suggestions? > > Matthias Jung wrote: > What about an automated shell script that downloads SystemC from > http://accellera.org/images/downloads/standards/systemc/systemc-2.3.1.tgz and > compiles it? Its maybe a little complicated but makes things easy for people > that don't have SystemC by default. Usually people install SystemC in > /opt/systemc so thats a place that could be checked for existance. > Christian? what do you think?
I don't like the idea of the download, and would rather suggest we should rely on pkg-config. Supposedly SystemC 2.3.1 actually does support pkg-config properly, with package names SystemC, TLM and TLM2. Could someone confirm? If that is the case, we could use the same flow as for protobuf for the determination of HAVE_SYSTEMC. The next step would be to build in support if SystemC is present and look in more detail at how we deal with "with python" and building of the gem5 lib. Perhaps we want to revisit this flow and not mandate going via the lib for SystemC (and rather just use a different main.cc)? - Andreas ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/3528/#review8436 ----------------------------------------------------------- On June 24, 2016, 10:20 a.m., Christian Menard wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/3528/ > ----------------------------------------------------------- > > (Updated June 24, 2016, 10:20 a.m.) > > > Review request for Default. > > > Repository: gem5 > > > Description > ------- > > The current TLM code only provides a Slave Port that allows the gem5 world to > send requests to the the TLM world. This patch adds a Master Port that allows > the TLM world to send requests to the gem5 world. Furthermore, the patch > provides a simple example application based on a TLM traffic generator. > > As of now the example code does not run error free. It eventually results in > an erro "fatal: Missed an event at time ...". So for now I leave the patch > for discussion until the error is resolved. > > > Diffs > ----- > > util/tlm/README dd6dfd38b6c2 > util/tlm/examples/master_port/Makefile PRE-CREATION > util/tlm/examples/master_port/main.cc PRE-CREATION > util/tlm/examples/master_port/tlm.py PRE-CREATION > util/tlm/examples/slave_port/Makefile PRE-CREATION > util/tlm/sc_master_port.hh PRE-CREATION > util/tlm/sc_master_port.cc PRE-CREATION > util/tlm/sim_control.cc PRE-CREATION > > Diff: http://reviews.gem5.org/r/3528/diff/ > > > Testing > ------- > > A simple example application consisting of a TLM traffic generator and a gem5 > memory. Currently the test fails. > > > Thanks, > > Christian Menard > > _______________________________________________ gem5-dev mailing list gem5-dev@gem5.org http://m5sim.org/mailman/listinfo/gem5-dev
