RFR: 8151670: Unexpected concurrent refinement deactivation and reactivation

Kim Barrett kim.barrett at oracle.com
Wed Mar 23 21:39:12 UTC 2016


> On Mar 23, 2016, at 9:00 AM, Mikael Gerdin <mikael.gerdin at oracle.com> wrote:
>> New webrev:
>> full: http://cr.openjdk.java.net/~kbarrett/8151670/webrev.01/
>> incr: http://cr.openjdk.java.net/~kbarrett/8151670/webrev.01.inc/
> 
> I'm ok with this version.
> 
> One thing, though.
> 178 inline void assert_fully_consumed(BufferNode* node, size_t buffer_size) {
> Moving assertions to inline functions is problematic because on Windows we cannot print the stack trace from an assertion context, so if this assert fires off on Windows there is no way to know which caller triggered it since the line number will always point to the inline function.
> 
> Would you be ok with making this a macro or passing in __LINE__ to assert_fully_consumed and print something like
> "Caller at " __FILE__ ":%d", line

Grr.  Thanks for the reminder.  I’ve macroized it.



More information about the hotspot-gc-dev mailing list