[UBSAN] Undefined behavior in Validation/*

[mrodozov]

The UBSAN IB reports undefined behavior in dqm Validation/* with example relval and step they appear in:

11603.0 step3
Validation/EcalDigis/src/EcalSelectiveReadoutValidation.cc:831:65: runtime error: member call on address 0x2adec9dd57e0 which does not point to an object of type 'EDProducerBase'

101.0 step1
Validation/EcalHits/src/EcalSimHitsValidProducer.cc:438:14: runtime error: index 26 out of bounds for type 'float [26]'

11603.0 step3
Validation/RecoTrack/plugins/MultiTrackValidator.cc:536:39: runtime error: variable length array bound evaluates to non-positive value 0

check the relval logs in here for the examples:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/ubsan_logs/relvals/

[mrodozov]

assign dqm

[cmsbuild]

New categories assigned: dqm

@jfernan2,@kmaeshima,@rvenditti,@andrius-k,@ErnestaP,@ahmad3213 you have been requested to review this Pull request/Issue and eventually sign? Thanks

[cmsbuild]

A new Issue was created by @mrodozov Mircho Rodozov.

@Dr15Jones, @perrotta, @dpiparo, @makortel, @smuzaffar, @qliphy can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[jfernan2]

Adding into the thread Tracker and ECAL DQM developers:
@sroychow @arossi83 @mmusich
@alejands @abhih1 @prebello @mbandrews

This issue looks related to: #35004

[mmusich]

@mrodozov @jfernan2
please have a look to #35086 for what concerns the MultiTrackValidator code.

[abhih1]

Please have a look at #35130 related to the Ecal Validation codes.

[mrodozov]

@mmusich there is another one now there, on a different line:
/data/cmsbld/jenkins/workspace/ib-run-pr-tests/CMSSW_12_1_UBSAN_X_2021-08-27-2300/src/Validation/RecoTrack/plugins/MultiTrackValidator.cc:486:27: runtime error: variable length array bound evaluates to non-positive value 0
see

https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-d54118/18157/runTheMatrix-results/11603.0_SingleElectronPt1000+2021+SingleElectronPt1000_pythia8_GenSim+Digi+Reco+HARVEST+ALCA/step3_SingleElectronPt1000+2021+SingleElectronPt1000_pythia8_GenSim+Digi+Reco+HARVEST+ALCA.log

there are few others but in FWCore

@abhih1 when the test is finished I'll check it

[mmusich]

/data/cmsbld/jenkins/workspace/ib-run-pr-tests/CMSSW_12_1_UBSAN_X_2021-08-27-2300/src/Validation/RecoTrack/plugins/MultiTrackValidator.cc:486:27: runtime error: variable length array bound evaluates to non-positive value 0

I know, unfortunately I won't be able to fix that one before next week.

[mmusich]

this can be signed and closed , right?

[smuzaffar]

@mmusich , there is still one error left in Validation/*

https://cmssdt.cern.ch/SDT/jenkins-artifacts/ubsan_logs/CMSSW_12_1_X_2021-10-08-2300/

281.0/step3:Validation/EcalDigis/plugins/EcalSelectiveReadoutValidation.cc:449:35: runtime error: index 54 out of bounds for type 'bool [54][68]'

[mmusich]

@abhih1 your PR #35130 has not been enough, please check #35037 (comment)

[abhih1]

@mmusich I have just submitted this PR for this error: #35613. Please have a look.

[smuzaffar]

closing it, as we do not get these runtime errors for UBSAN builds

[cmsbuild]

cms-bot internal usage