JDK-8144573 : TLABWasteIncrement=max_jint fires an assert on SPARC for non-G1 GC mode
  • Type: Bug
  • Component: hotspot
  • Sub-Component: gc
  • Affected Version: 9
  • Priority: P2
  • Status: Resolved
  • Resolution: Fixed
  • Submitted: 2015-12-03
  • Updated: 2016-02-11
  • Resolved: 2016-01-11
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 9
9 b105Fixed
Related Reports
Relates :  
JDK-8142341 - GC: current flags need ranges to be implemented
Relates :  
JDK-8144578 - TestOptionsWithRanges test only ever uses the default collector
Description
TLABWasteIncrement=max_intx is okay for G1 on SPARC but other collectors fires an assert.
ThreadLocalAllocBuffer::refill_waste_limit_increment() is directly used at 'add' instruction but only 13bits are used on that.

In addition, as current TestOptionsWithRanges.java tests only for default garbage collector if there is no dependency on the flag, we couldn't find this issue before. 
With JDK-8144578 fixed, we don't need a new test for all collectors.
Comments
As G1 only has false return from Universe::heap()->supports_inline_contig_alloc(), G1 doesn't reach to the problematic instruction which uses ThreadLocalAllocBuffer::refill_waste_limit_increment() at TemplateTable::_new()
16-12-2015
Right, we don't need a new test for this issue with JDK-8144578 fixed. I updated the description.
07-12-2015
Probably, new test is not needed for that problem. This flag will be tested for non-G1 GC mode after fixing JDK-8144578.
07-12-2015
Attaching hs_err_pid13562.log when we hit the assert
03-12-2015