jdk8-mips64-public/hotspot: changeset 1052:0db4adb6e914

6810698: G1: two small bugs in the sparse remembered sets
Summary: The _expanded flag of the sparse RSets is not reset and this can leave a RSet in an inconsistent state if it is expanded more than once. Also, we should be iterating over the _cur, instead of the _next, sparse table
Reviewed-by: apetrusenko, iveresov

     1.1 --- a/src/share/vm/gc_implementation/g1/sparsePRT.cpp	Fri Mar 06 13:50:14 2009 -0800
     1.2 +++ b/src/share/vm/gc_implementation/g1/sparsePRT.cpp	Sat Mar 07 11:07:36 2009 -0500
     1.3 @@ -504,6 +504,7 @@
     1.4    // Make sure that the current and next tables agree.  (Another mechanism
     1.5    // takes care of deleting now-unused tables.)
     1.6    _cur = _next;
     1.7 +  set_expanded(false);
     1.8  }
     1.9  
    1.10  void SparsePRT::expand() {

     2.1 --- a/src/share/vm/gc_implementation/g1/sparsePRT.hpp	Fri Mar 06 13:50:14 2009 -0800
     2.2 +++ b/src/share/vm/gc_implementation/g1/sparsePRT.hpp	Sat Mar 07 11:07:36 2009 -0500
     2.3 @@ -274,7 +274,7 @@
     2.4  
     2.5    // Clean up all tables on the expanded list.  Called single threaded.
     2.6    static void cleanup_all();
     2.7 -  RSHashTable* next() const { return _next; }
     2.8 +  RSHashTable* cur() const { return _cur; }
     2.9  
    2.10  
    2.11    void init_iterator(SparsePRTIter* sprt_iter);
    2.12 @@ -300,7 +300,7 @@
    2.13    {}
    2.14  
    2.15    void init(const SparsePRT* sprt) {
    2.16 -    RSHashTableIter::init(sprt->next());
    2.17 +    RSHashTableIter::init(sprt->cur());
    2.18    }
    2.19    bool has_next(size_t& card_index) {
    2.20      return RSHashTableIter::has_next(card_index);

src/share/vm/gc_implementation/g1/sparsePRT.cpp		file \| annotate \| diff \| comparison \| revisions
src/share/vm/gc_implementation/g1/sparsePRT.hpp		file \| annotate \| diff \| comparison \| revisions

Mercurial > jdk8-mips64-public > hotspot / changeset

changeset

6810698: G1: two small bugs in the sparse remembered sets