-
Notifications
You must be signed in to change notification settings - Fork 4.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[MkFit] Fix UBSAN error - skip processing of layers with no hits. #37146
[MkFit] Fix UBSAN error - skip processing of layers with no hits. #37146
Conversation
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-37146/28688
|
A new Pull Request was created by @osschar (Matevž Tadel) for master. It involves the following packages:
@jpata, @cmsbuild, @clacaputo, @slava77 can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
test parameters:
|
@cmsbuild, please test |
@smuzaffar this has been running tests for 3 days (!?). |
@mmusich , UBSAN IB comparison does not work. The job timedout after 16 hours ( https://cmssdt.cern.ch/jenkins/job/compare-root-files-short-matrix/48727/ ) . Check the relvals for wf 10004.0 to see if USBAN issue is fixed or not? I have restarted the comparison job now but I am afraid it will timeout again. |
please test |
+1 Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-cdca02/22911/summary.html Comparison Summary@slava77 comparisons for the following workflows were not done due to missing matrix map:
Summary:
|
Hi @smuzaffar, where can I check wf 10004.0 relvals for UBSAN? |
URL (i.e. https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-cdca02/22856) for |
-1 Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-cdca02/22856/summary.html |
ok, I have forced set the status for |
If I check the output in
|
I observe the same locally: cmsrel CMSSW_12_3_UBSAN_X_2022-03-04-1100
cd CMSSW_12_3_UBSAN_X_2022-03-04-1100/src/
cmsenv
git cms-merge-topic 37146
scramv1 b -j 20
runTheMatrix.py -l 10004.0 -t 4 -j 8
more 10004.0_SingleGammaPt10+2017+SingleGammaPt10_pythia8_GenSim+Digi+RecoFakeHLT+HARVESTFakeHLT+ALCA+Nano/step3_SingleGammaPt10+2017+SingleGammaPt10_pythia8_GenSim+Digi+RecoFakeHLT+HARVESTFakeHLT+ALCA+Nano.log | grep libRecoTrackerMkFitCore | wc -l
2 |
Oh rats. How does one setup the build environment for this? |
…eeds Shortcut MkFitProducer to produce an empty track vector when input seed vector is empty. In MkBuilder::import_seeds() assert that the input vector is not empty.
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-37146/28767
|
Pull request #37146 was updated. @jpata, @cmsbuild, @clacaputo, @slava77 can you please check and sign again. |
@cmsbuild please test |
+1 Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-cdca02/22994/summary.html Comparison Summary@slava77 comparisons for the following workflows were not done due to missing matrix map:
Summary:
|
+reconstruction
cmsrel CMSSW_12_3_UBSAN_X_2022-03-04-1100
cd CMSSW_12_3_UBSAN_X_2022-03-04-1100/src/
cmsenv
git cms-merge-topic 37146
scramv1 b -j 20
runTheMatrix.py -l 10004.0 -t 4 -j 8
more 10004.0_SingleGammaPt10+2017+SingleGammaPt10_pythia8_GenSim+Digi+RecoFakeHLT+HARVESTFakeHLT+ALCA+Nano/step3_SingleGammaPt10+2017+SingleGammaPt10_pythia8_GenSim+Digi+RecoFakeHLT+HARVESTFakeHLT+ALCA+Nano.log | grep libRecoTrackerMkFitCore | wc -l
0
|
This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @perrotta, @dpiparo, @qliphy (and backports should be raised in the release meeting by the corresponding L2) |
+1 |
I suppose it would be useful (to minimize UB in 12_3). |
…-1100_mkFit_X [MkFit] Fix UBSAN error in (almost) empty events (backport of #37146)
When processing layers with no hits the LayerOfHits vector was in default state with storage pointer set to nullptr.
Loaders of hits for vector processing were still being initialized, resulting in doing pointer arithmetic with nullptr -- though they were never actually used.
This should fix #37136. Note that the wf there was 10004.0, single photon 10GeV and there were indeed several occurrences of empty mkfit hit containers.