-
Notifications
You must be signed in to change notification settings - Fork 4.3k
-
Notifications
You must be signed in to change notification settings - Fork 4.3k
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
[ARM] Crashes in Cling #43802
Comments
cms-bot internal usage |
A new Issue was created by @makortel Matti Kortelainen. @rappoccio, @sextonkennedy, @Dr15Jones, @smuzaffar, @antoniovilela, @makortel can you please review it and eventually sign/assign? Thanks. cms-bot commands are listed here |
Some more from CMSSW_14_0_X_2024-01-25-2300 on el9_aarch64_gcc12 140.58 step 3
11024.0 step 3
25034.114 step 3
|
assign core |
New categories assigned: core @Dr15Jones,@makortel,@smuzaffar you have been requested to review this Pull request/Issue and eventually sign? Thanks |
type root |
Responding to @vgvassilev's #43577 (comment)
I ran valgrind for step3 of workflow 11025.0 in CMSSW_14_0_X_2024-01-11-2300. Below is a summary of different kinds of warnings
Similar report repeats for
Repeats for
|
More from the same valgrind run
|
So I didn't see anything obvious related to Cling, but we might want to follow up
|
The following errors usually indicates uninitialised memory in the user data.
and, at least in the first case, valgrind points to:
|
is a bit worrisome and might be an issue in ROOT (or valgrind) |
Thanks for confirming, this was my interpretation as well. |
The following is the most unexpected of all and may indicate either that valgrind is not ready for the platform or that there is a serious memory corruption somewhere. Since the symbol mentions LTO, did you also try without LTO turned on.
|
Ok, there were more of those. I attached a file that should contain all of them (in case that would be helpful). |
I did not (and at this stage I'd have to move to newer IB). @smuzaffar Is it possible to turn off LTO for select package(s) in a local build, or would that require using the NONLTO build? |
@makortel , currently you can not disable LTO for individual packages. But what you can do is to
this should allow you to build the checked out packages without LTO |
Hello, I am not 100% sure it is fully related, but workflow
Log from today's IBs at CMSSW_14_1_X_2024-02-26-2300. RelVal
|
Two new crashes in CMSSW_14_1_ROOT6_X_2024-03-06-2300:
and 11025.0
|
Any chance to cook up a ROOT-only reproducer? |
Crash in CMSSW_14_1_X_2024-05-28-1100
|
Hello, We see multiple of those crashed in cling reporting two types of stack trace:
|
This issue is to continue to track the crashes in Cling we're seeing on ARM that were first reported in #43577 (comment). Repeating the stack traces:
On el8_aarch64_gcc12 CMSSW_14_0_X_2023-12-13-2300 two workflows crashed in a way that looks possibly related
11024.0 step 3
https://cmssdt.cern.ch/SDT/cgi-bin/logreader/el8_aarch64_gcc12/CMSSW_14_0_X_2023-12-13-2300/pyRelValMatrixLogs/run/11024.0_TTbar_13+2018PU/step3_TTbar_13+2018PU.log#/
11025.0 step 3
https://cmssdt.cern.ch/SDT/cgi-bin/logreader/el8_aarch64_gcc12/CMSSW_14_0_X_2023-12-13-2300/pyRelValMatrixLogs/run/11025.0_ZEE_13+2018PU/step3_ZEE_13+2018PU.log#/
The text was updated successfully, but these errors were encountered: