RFR(s): backport of 8031126: java/lang/management/ThreadMXBean/ThreadUserTime.java fails intermittently

[Fedor Burdun]

Hello everyone,

The issue described in bugs 8031126/8030631 is still reproducible on java8.
May I request a backport of the changeset http://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/af53a220ea60 to java8?

The adjusted patch is:


<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

diff --git a/src/os/linux/vm/os_linux.cpp b/src/os/linux/vm/os_linux.cpp
--- a/src/os/linux/vm/os_linux.cpp
+++ b/src/os/linux/vm/os_linux.cpp
@@ -6093,8 +6093,6 @@
 PRAGMA_DIAG_PUSH
 PRAGMA_FORMAT_NONLITERAL_IGNORED
 static jlong slow_thread_cpu_time(Thread *thread, bool user_sys_cpu_time) {
-  static bool proc_task_unchecked = true;
-  static const char *proc_stat_path = "/proc/%d/stat";
   pid_t  tid = thread->osthread()->thread_id();
   char *s;
   char stat[2048];
@@ -6107,23 +6105,7 @@
   long ldummy;
   FILE *fp;

-  // The /proc/<tid>/stat aggregates per-process usage on
-  // new Linux kernels 2.6+ where NPTL is supported.
-  // The /proc/self/task/<tid>/stat still has the per-thread usage.
-  // See bug 6328462.
-  // There possibly can be cases where there is no directory
-  // /proc/self/task, so we check its availability.
-  if (proc_task_unchecked && os::Linux::is_NPTL()) {
-    // This is executed only once
-    proc_task_unchecked = false;
-    fp = fopen("/proc/self/task", "r");
-    if (fp != NULL) {
-      proc_stat_path = "/proc/self/task/%d/stat";
-      fclose(fp);
-    }
-  }
-
-  sprintf(proc_name, proc_stat_path, tid);
+  snprintf(proc_name, 64, "/proc/self/task/%d/stat", tid);
   fp = fopen(proc_name, "r");
   if ( fp == NULL ) return -1;
   statlen = fread(stat, 1, 2047, fp);
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<


Unfortunately I can not update OpenJDK bug for such request since I have not OpenJDK id.

Thanks,
Fedor