Michael Ho has posted comments on this change. ( http://gerrit.cloudera.org:8080/12545 )
Change subject: KUDU-2706: Work around the lack of thread safety in krb5_parse_name() ...................................................................... Patch Set 2: (3 comments) http://gerrit.cloudera.org:8080/#/c/12545/1//COMMIT_MSG Commit Message: http://gerrit.cloudera.org:8080/#/c/12545/1//COMMIT_MSG@17 PS1, Line 17: In addition, krb5_parse_name() is not thread safe as > I think it's clearer to explain that krb5_parse_name isn't thread-safe beca Done http://gerrit.cloudera.org:8080/#/c/12545/1//COMMIT_MSG@20 PS1, Line 20: tialization may be inadvertently modified concu > BTW, krb5_set_error_message() is also not thread safe. However, if you are Good point. That said, based on the stack trace in KUDU-2706, the workaround seems sufficient. However, as you pointed out, in case we hit any error in Kerberos library, krb5_set_error_message() may lead to crash due to unsafe sharing of 'g_krb5_ctx'. I cannot think of a easier workaround other than addressing the unsafe use of 'g_krb5_ctx' directly. http://gerrit.cloudera.org:8080/#/c/12545/1//COMMIT_MSG@22 PS1, Line 22: > it's not really explicit -- you're implicitly initializing it by forcing kr Removed the word 'explicit'. -- To view, visit http://gerrit.cloudera.org:8080/12545 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: kudu Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I1bf9224516e2996f51f319088179727f76741ebe Gerrit-Change-Number: 12545 Gerrit-PatchSet: 2 Gerrit-Owner: Michael Ho <[email protected]> Gerrit-Reviewer: Alexey Serbin <[email protected]> Gerrit-Reviewer: Kudu Jenkins (120) Gerrit-Reviewer: Michael Ho <[email protected]> Gerrit-Reviewer: Todd Lipcon <[email protected]> Gerrit-Comment-Date: Thu, 21 Feb 2019 19:48:31 +0000 Gerrit-HasComments: Yes
