archive/jdk
1
0
Fork 0
mirror of https://github.com/openjdk/jdk.git synced 2025-08-26 06:14:49 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

7023069: G1: Introduce symmetric locking in the slow allocation path

Browse source 
7023151: G1: refactor the code that operates on _cur_alloc_region to be re-used for allocs by the GC threads
7018286: G1: humongous allocation attempts should take the GC locker into account

First, this change replaces the asymmetric locking scheme in the G1 slow alloc path by a summetric one. Second, it factors out the code that operates on _cur_alloc_region so that it can be re-used for allocations by the GC threads in the future.

Reviewed-by: stefank, brutisso, johnc
This commit is contained in:
Antonios Printezis 2011-03-30 10:26:59 -04:00
parent 349d820dd1
commit 3e9fe24ddd
11 changed files with 920 additions and 747 deletions
Download patch file Download diff file Expand all files Collapse all files

11
hotspot/src/share/vm/gc_implementation/g1/heapRegion.cpp
View file

@ -360,6 +360,7 @@ void HeapRegion::hr_clear(bool par, bool clear_space) {
set_young_index_in_cset(-1);
uninstall_surv_rate_group();
set_young_type(NotYoung);
reset_pre_dummy_top();
if (!par) {
// If this is parallel, this will be done later.
@ -923,11 +924,11 @@ void G1OffsetTableContigSpace::set_saved_mark() {
ContiguousSpace::set_saved_mark();
OrderAccess::storestore();
_gc_time_stamp = curr_gc_time_stamp;
// The following fence is to force a flush of the writes above, but
// is strictly not needed because when an allocating worker thread
// calls set_saved_mark() it does so under the ParGCRareEvent_lock;
// when the lock is released, the write will be flushed.
// OrderAccess::fence();
// No need to do another barrier to flush the writes above. If
// this is called in parallel with other threads trying to
// allocate into the region, the caller should call this while
// holding a lock and when the lock is released the writes will be
// flushed.
}
}