Potential memory leak in FastSim TrackerSimHitProducer

[kpedro88]

Reported by valgrind:

cmssw/FastSimulation/SimplifiedGeometryPropagator/plugins/TrackerSimHitProducer.cc

Line 311 in f347d6f

new PSimHit(entry, exit, localMomentum.mag(), tof, eLoss, pdgId,

The PSimHit returned by fastsim::TrackerSimHitProducer::createHitOnDetector() is allocated with new. It gets stored in a map here:

cmssw/FastSimulation/SimplifiedGeometryPropagator/plugins/TrackerSimHitProducer.cc

Line 220 in f347d6f

std::pair<double, PSimHit*> hitPair = createHitOnDetector(particleState, pdgId, layer.getThickness(particle.position()), energyDeposit, simTrackId, detector, positionOutside);

And then copied by value into the container:

cmssw/FastSimulation/SimplifiedGeometryPropagator/plugins/TrackerSimHitProducer.cc

Line 240 in f347d6f

simHitContainer_->push_back(*(it->second));

This implies the original new'd object is never deleted. It would be better to move around the PSimHit using smart pointers.

[kpedro88]

assign fastsim

[cmsbuild]

New categories assigned: fastsim

@mdhildreth,@ssekmen,@lveldere,@civanch you have been requested to review this Pull request/Issue and eventually sign? Thanks

[cmsbuild]

A new Issue was created by @kpedro88 Kevin Pedro.

@davidlange6, @Dr15Jones, @smuzaffar, @fabiocos, @kpedro88 can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[kpedro88]

@ssekmen @skurz this should be addressed before the 2017 fastsim workflow is put into production

[kpedro88]

addressed by #24210