RFR [XXS] 8225402: events logging in deoptimization.cpp should go to deopt-log

coleen.phillimore at oracle.com coleen.phillimore at oracle.com
Thu Jun 6 12:03:17 UTC 2019


+1 Thank you!
Coleen

On 6/6/19 5:07 AM, Thomas Stüfe wrote:
> +1
>
> ..Thomas
>
> On Thu, Jun 6, 2019 at 10:22 AM Baesken, Matthias <matthias.baesken at sap.com>
> wrote:
>
>>> Got it. Yes, I agree that they should be also in deopt log.
>>>
>> Hello here is an XXS change that  adjusts the  events - logging in
>> deoptimization.cpp  as discussed . Please review .
>>
>>
>> Bug/webrev :
>>
>> https://bugs.openjdk.java.net/browse/JDK-8225402
>>
>> http://cr.openjdk.java.net/~mbaesken/webrevs/8225402.0/
>>
>>
>> Thanks, Matthias
>>
>>
>>
>>> -----Original Message-----
>>> From: Vladimir Kozlov <vladimir.kozlov at oracle.com>
>>> Sent: Mittwoch, 5. Juni 2019 18:40
>>> To: Baesken, Matthias <matthias.baesken at sap.com>; hotspot-
>>> dev at openjdk.java.net
>>> Subject: Re: deoptimization.cpp Events::log / Events::log_deopt_message -
>>> was : RE: RFR: 8224221: add memprotect calls to event log
>>>
>>> Got it. Yes, I agree that they should be also in deopt log.
>>>
>>> Thanks,
>>> Vladimir
>>>
>>> On 6/4/19 1:27 AM, Baesken, Matthias wrote:
>>>> Hi Vladimir ,  sure  the  log_deopt_message() is used  for the
>> event-logging
>>> and the events are later shown  in the hs_err file  .
>>>> However my question was :  in  deoptimization.cpp   we have  ***both***
>>> calls  to   Events::log   and     Events::log_deopt_message   .
>>>>     // Log a message
>>>>     Events::log(thread, "Uncommon trap: trap_request=" PTR32_FORMAT "
>>> fr.pc=" INTPTR_FORMAT " relative=" INTPTR_FORMAT,
>>>>                 trap_request, p2i(fr.pc()), fr.pc() -
>> fr.cb()->code_begin());
>>>> Events::log(thread, "DEOPT PACKING pc=" INTPTR_FORMAT " sp="
>>> INTPTR_FORMAT, p2i(fr.pc()), p2i(fr.sp()));
>>>>
>>>>       // Log a message
>>>>       Events::log_deopt_message(thread, "Uncommon trap: reason=%s
>>> action=%s pc=" INTPTR_FORMAT " method=%s @ %d %s",
>>>>                                 trap_reason_name(reason),
>> trap_action_name(action),
>>> p2i(fr.pc()),
>> trap_method->name_and_sig_as_C_string(), trap_bci, nm-
>>>> compiler_name());
>>>>
>>>>
>>>>
>>>>
>>>> I think they  all  should  call  Events::log_deopt_message ,  so that
>> they go
>>> into the  same event log part  that is collecting the de-opt  event
>> messages .
>>



More information about the hotspot-dev mailing list