archive/jdk
1
0
Fork 0
mirror of https://github.com/openjdk/jdk.git synced 2025-09-20 19:14:38 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

8028126: nsk/jvmti/scenarios/hotswap/HS101/hs101t006 Crashed the vm on Solaris-sparc64 fastdebug builds: only current thread can flush its registers

Browse source 
Fix a race between VMOp_GetCurrentLocation reaching a safepoint and arget thread exiting from Java execution

Reviewed-by: sla, dholmes, dsamersoff
This commit is contained in:
Serguei Spitsyn 2013-12-03 15:41:35 -08:00 • committed by Serguei Spitsyn
parent b5f2f33aaa
commit e820946a1e
1 changed files with 14 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

19
hotspot/src/share/vm/prims/jvmtiEnvThreadState.cpp
View file

@ -269,11 +269,20 @@ class VM_GetCurrentLocation : public VM_Operation {
void doit() {
ResourceMark rmark; // _thread != Thread::current()
RegisterMap rm(_thread, false);
javaVFrame* vf = _thread->last_java_vframe(&rm);
assert(vf != NULL, "must have last java frame");
Method* method = vf->method();
_method_id = method->jmethod_id();
_bci = vf->bci();
// There can be a race condition between a VM_Operation reaching a safepoint
// and the target thread exiting from Java execution.
// We must recheck the last Java frame still exists.
if (_thread->has_last_Java_frame()) {
javaVFrame* vf = _thread->last_java_vframe(&rm);
assert(vf != NULL, "must have last java frame");
Method* method = vf->method();
_method_id = method->jmethod_id();
_bci = vf->bci();
} else {
// Clear current location as the target thread has no Java frames anymore.
_method_id = (jmethodID)NULL;
_bci = 0;
}
}
void get_current_location(jmethodID *method_id, int *bci) {
*method_id = _method_id;