> On Sept. 24, 2015, 11:03 p.m., Steve Reinhardt wrote: > > Looks fine to me. I assume Nate wrote all this code originally, so it > > would be interesting to hear from him, but no need to wait for that to > > happen. > > > > What is the use case for having the global enable/disable flag anyway? Is > > this how --trace-start is implemented? Other than that I can't even think > > of a use for it.
Yes, --debug-start <tick> uses this mechanism. - Curtis ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/3082/#review7276 ----------------------------------------------------------- On Aug. 31, 2015, 6:35 p.m., Curtis Dunham wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/3082/ > ----------------------------------------------------------- > > (Updated Aug. 31, 2015, 6:35 p.m.) > > > Review request for Default. > > > Repository: gem5 > > > Description > ------- > > The DTRACE() macro tests both Trace::enabled and the specific flag. This > change uses the same administrative interface for enabling/disabling > tracing, but masks the SimpleFlags settings directly. This eliminates a > load for every DTRACE() test, e.g. DPRINTF. > > This functionality was not and remains not threadsafe. > > > Diffs > ----- > > src/base/trace.cc 969113566d5056fcbccc555e74aa0eb14fae4398 > src/cpu/exetrace.hh 969113566d5056fcbccc555e74aa0eb14fae4398 > src/cpu/inst_pb_trace.cc 969113566d5056fcbccc555e74aa0eb14fae4398 > src/cpu/inteltrace.hh 969113566d5056fcbccc555e74aa0eb14fae4398 > src/python/m5/trace.py 969113566d5056fcbccc555e74aa0eb14fae4398 > src/python/swig/trace.i 969113566d5056fcbccc555e74aa0eb14fae4398 > src/base/debug.hh 969113566d5056fcbccc555e74aa0eb14fae4398 > src/base/debug.cc 969113566d5056fcbccc555e74aa0eb14fae4398 > src/base/trace.hh 969113566d5056fcbccc555e74aa0eb14fae4398 > > Diff: http://reviews.gem5.org/r/3082/diff/ > > > Testing > ------- > > On ARM quick regressions, runtime reductions of up to 18% were observed. > > > Thanks, > > Curtis Dunham > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
