archive/jdk
1
0
Fork 0
mirror of https://github.com/openjdk/jdk.git synced 2025-08-26 22:34:27 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

6988458: G1: assert(mr.end() <= _cm->finger()) failed: otherwise the region shouldn't be on the stack

Browse source 
The changes from 6941395 did not clear the CMTask::_aborted_region fields when concurrent marking aborted because of overflow. As a result, the next time around we could see a memory region whose start address was above the global finger and the assertion tripped. Moved the clearing of the aborted regions to ConcurrentMark::clear_marking_state, which is executed on all of the exit paths.

Reviewed-by: tonyp, ysr, jmasa
This commit is contained in:
John Cuthbertson 2010-10-18 15:01:41 -07:00
parent 0c9bfb6003
commit f6acb9efb0
1 changed files with 3 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

5
hotspot/src/share/vm/gc_implementation/g1/concurrentMark.cpp
View file

@ -2418,6 +2418,8 @@ void ConcurrentMark::clear_marking_state() {
for (int i = 0; i < (int)_max_task_num; ++i) {
OopTaskQueue* queue = _task_queues->queue(i);
queue->set_empty();
// Clear any partial regions from the CMTasks
_tasks[i]->clear_aborted_region();
}
}
@ -2706,7 +2708,6 @@ void ConcurrentMark::abort() {
clear_marking_state();
for (int i = 0; i < (int)_max_task_num; ++i) {
_tasks[i]->clear_region_fields();
_tasks[i]->clear_aborted_region();
}
_has_aborted = true;
@ -2985,7 +2986,7 @@ void CMTask::reset(CMBitMap* nextMarkBitMap) {
_nextMarkBitMap = nextMarkBitMap;
clear_region_fields();
clear_aborted_region();
assert(_aborted_region.is_empty(), "should have been cleared");
_calls = 0;
_elapsed_time_ms = 0.0;