stan/jdk-24
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

6791815: Fix for 6471657 can cause deadlock on non-Solaris platforms when initializing direct buffer support

Browse Source 
Place the state transition inside the loop so that the VMThread could proceed for safepoint

Reviewed-by: dholmes, never, acorn
This commit is contained in:
Xiaobin Lu 2009-01-13 12:08:43 -08:00
parent 9bf4545843
commit aae30d753f
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
hotspot/src/share/vm/prims/jni.cpp
View File

@ -2691,8 +2691,13 @@ static bool initializeDirectBufferSupport(JNIEnv* env, JavaThread* thread) {
directBufferSupportInitializeEnded = 1;
} else {
ThreadInVMfromNative tivn(thread); // set state as yield_all can call os:sleep
while (!directBufferSupportInitializeEnded && !directBufferSupportInitializeFailed) {
// Set state as yield_all can call os:sleep. On Solaris, yield_all calls
// os::sleep which requires the VM state transition. On other platforms, it
// is not necessary. The following call to change the VM state is purposely
// put inside the loop to avoid potential deadlock when multiple threads
// try to call this method. See 6791815 for more details.
ThreadInVMfromNative tivn(thread);
os::yield_all();
}
}