| Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
|---|---|---|---|---|---|---|
| JDK-2208963 | 7 | Tony Printezis | P4 | Closed | Fixed | b138 |
Currently we have two versions of code that attempts to satisfy parallel allocation requests out of a heap region and attempts to replace said heap region when it's full:
a) One version is for allocations by mutator threads.
b) Another version is for allocations by GC threads (i.e., when they need to copy objects from from- to to-space).
It'd be helpful to consolidate these two versions and have a single version of the code shared by the two paths above. This CR proposes the refactoring of the code of path a) in a separate class (G1AllocRegion). This is one step towards sharing this code by path b) (this will be done as a separate CR in the future).
a) One version is for allocations by mutator threads.
b) Another version is for allocations by GC threads (i.e., when they need to copy objects from from- to to-space).
It'd be helpful to consolidate these two versions and have a single version of the code shared by the two paths above. This CR proposes the refactoring of the code of path a) in a separate class (G1AllocRegion). This is one step towards sharing this code by path b) (this will be done as a separate CR in the future).
- backported by
-
JDK-2208963 G1: refactor the code that operates on _cur_alloc_region to be re-used for allocs by the GC threads
-
- Closed
-
- relates to
-
JDK-7033292 G1: nightly failure: Non-dirty cards in region that should be dirty
-
- Closed
-
-
JDK-7039627 G1: avoid BOT updates for survivor allocations and dirty survivor regions incrementally
-
- Closed
-