RFR(S): 8010738: G1: Output for full GCs with +PrintGCDetails should contain perm gen size/meta data change info

[Bengt Rutisson]

On 5/9/13 1:42 AM, John Cuthbertson wrote:
> Hi Bengt,
>
> Thanks for looking at the changes. Reply inline....
>
> On 5/8/2013 9:18 AM, Bengt Rutisson wrote:
>>
>> John, I just noticed one thing that I find a bit odd. The 
>> G1CollectorPolicy::print_detailed_heap_transition() method is only 
>> called from G1CollectedHeap::do_collection(), so it really doesn't 
>> have to take the full parameter. It could always print the metaspace 
>> information, right?
>
> There are two calls to 
> G1CollectorPolicy::print_detailed_heap_transition(). The other is from 
> G1CollectedHeap::log_gc_footer():

Ah! I missed that since I was only looking through the webrev (on my 
phone) and there is a default value for the full parameter, so there was 
only one place where print_detailed_heap_transition() had been updated. 
Sorry for the confusion.

Bengt


>
>> void G1CollectedHeap::log_gc_footer(double pause_time_sec) {
>>   if (!G1Log::fine()) {
>>     return;
>>   }
>>
>>   if (G1Log::finer()) {
>>     if (evacuation_failed()) {
>>       gclog_or_tty->print(" (to-space exhausted)");
>>     }
>>     gclog_or_tty->print_cr(", %3.7f secs]", pause_time_sec);
>>     g1_policy()->phase_times()->note_gc_end();
>>     g1_policy()->phase_times()->print(pause_time_sec);
>>     g1_policy()->print_detailed_heap_transition();
>>   } else {
>>     if (evacuation_failed()) {
>>       gclog_or_tty->print("--");
>>     }
>>     g1_policy()->print_heap_transition();
>>     gclog_or_tty->print_cr(", %3.7f secs]", pause_time_sec);
>>   }
>>   gclog_or_tty->flush();
>> }
>
> and comes at the end of an evacuation pause. So I think we need the 
> parameter - unless we decide to include in the output for each GC.
>
> JohnC

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20130510/cd372083/attachment.htm>