Matti Kortelainen

Occurred in * CMSSW_14_1_X_2024-03-16-1100 on slc7_amd64_gcc12 wf [140.004 step 3](https://cmssdt.cern.ch/SDT/cgi-bin/logreader/slc7_amd64_gcc12/CMSSW_14_1_X_2024-03-16-1100/pyRelValMatrixLogs/run/140.004_RunBTagMu2022A/step3_RunBTagMu2022A.log#/) * CMSSW_14_1_NONLTO_X_2024-03-16-1100 wf [140.021 step 3](https://cmssdt.cern.ch/SDT/cgi-bin/logreader/el8_amd64_gcc12/CMSSW_14_1_NONLTO_X_2024-03-16-1100/pyRelValMatrixLogs/run/140.021_RunMinimumBias2022B/step3_RunMinimumBias2022B.log#/) In CMSSW_14_1_NONLTO_X_2024-03-16-1100 wf 140.009 step 3 showed new kind of stack trace ``` Thread...

The stack traces from CLANG and latest from NONLTO with `std::_Rb_tree` hint towards https://github.com/cms-sw/cmssw/blob/7827a0793b6e40d17acc29b32b5e2f488b0c8415/DPGAnalysis/Skims/src/LogErrorEventFilter.cc#L205-L209

I ran valgrind on the step3 of 140.009, that showed ``` ==24701== Thread 4: ==24701== Conditional jump or move depends on uninitialised value(s) ==24701== at 0x6A5E3362: .LTHUNK17.lto_priv.0 (in /cvmfs/cms-ib.cern.ch/sw/x86_64/nweek-02828/el9_amd64_gcc12/cms/cmssw/CMSSW_14_1_X_2024-03-14-2300/lib/el9_amd64_gcc12/pluginDPGAnalysisSkims.so) ==24701==...

> the really obvious question, which I decided not to ask earlier: why are they coding their own spinlock? It was done in https://github.com/cms-sw/cmssw/pull/22329, so the question is more "why...

My valgrind showed only https://github.com/cms-sw/cmssw/issues/44413#issuecomment-2003282758 as relevant to this issue.

assign xpog

+core Although in the end there wasn't much (anything?) for core.

@cmsbuild, please close

@cmsbuild, please close

unhold