Patch for undefined behaviour (nullpointer reference).#26
Open
ursg wants to merge 1 commit intofmihpc:vlasiator-versionfrom
Open
Patch for undefined behaviour (nullpointer reference).#26ursg wants to merge 1 commit intofmihpc:vlasiator-versionfrom
ursg wants to merge 1 commit intofmihpc:vlasiator-versionfrom
Conversation
Discovered by GCC12's undefined behaviour sanitizer.
|
As commented above that code, this can also be fixed by using temp_result.data() instead of &(temp_result[0]) and by now c++11 is old enough for that :) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
For communication of empty (zero velocity blocks) Vlasiator cells, this was trying to do an AllGather of some empty arrays, but the actual Allgatherv call tried to acess
&(temp_result[0]), which is technically a nullpointer dereference.This probably wasn't causing any actual issue (because every sane compiler would optimize that away anyway), but was reported by GCC12's undefined behaviour sanitizer.
But anyway, if total_send_count is zero in the AllGather, the whole second stage of the communication can anyway be skipped, so that's how this is fixed now.