[ https://issues.apache.org/jira/browse/HDFS-15270?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17099990#comment-17099990 ]
Babneet Singh commented on HDFS-15270: -------------------------------------- The null check is missing in [every release branch|https://hadoop.apache.org/docs/current/hadoop-project-dist/hadoop-common/release/index.html]. [~weichiu] Do we backport/cherrypick to only those release branches, which are relevant to the issue's originator? > Account for *env == NULL in hdfsThreadDestructor > ------------------------------------------------ > > Key: HDFS-15270 > URL: https://issues.apache.org/jira/browse/HDFS-15270 > Project: Hadoop HDFS > Issue Type: Bug > Environment: Please refer to the "steps to reproduce" the failure inĀ > https://github.com/eclipse/openj9/issues/7752#issue-521732953. > Reporter: Babneet Singh > Assignee: Babneet Singh > Priority: Major > Fix For: 3.4.0 > > > OpenJ9 JVM properly terminates the thread before hdfsThreadDestructor is > invoked. JNIEnv is a mirror of J9VMThread in OpenJ9. After proper thread > termination, accessing JNIEnv in hdfsThreadDestructor (*env)->GetJavaVM, > yields a SIGSEGV since *env is NULL after thread cleanup is performed. > The main purpose of hdfsThreadDestructor is to invoke > DetachCurrentThread, which performs thread cleanup in OpenJ9. Since > OpenJ9 performs thread cleanup before hdfsThreadDestructor is invoked, > hdfsThreadDestructor should account for *env == NULL and skip > DetachCurrentThread. -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org