RFR(M): 8007439 : C2: adding successful message of inlining

Vladimir Kozlov vladimir.kozlov at oracle.com
Thu Feb 21 08:08:33 PST 2013


Good.

Thanks,
Vladimir

On 2/21/13 6:34 AM, Vladimir Ivanov wrote:
> Vladimir K.,
>
> LogCompilation tool works fine with these changes.
> I'll push the patch today.
>
> Best regards,
> Vladimir Ivanov
>
> On 2/13/13 12:51 AM, Vladimir Kozlov wrote:
>> I always forgetting about "official" part. Yes, one should be enough for
>> these changes. I wanted you to run our LogCompilation parser/tool to see
>> if it works with these changes before the push.
>>
>> Thanks,
>> Vladimir
>>
>> On 2/12/13 12:30 PM, Vladimir Ivanov wrote:
>>> Vladimir K.,
>>>
>>> The fix looks good to me.
>>>
>>> I think Igor is asking whether 1 "official" reviewer is enough for this
>>> change.
>>>
>>> Best regards,
>>> Vladimir Ivanov
>>>
>>> On 2/12/13 8:40 PM, Vladimir Kozlov wrote:
>>>> I assume Vladimir Ivanov should look on it since he is your sponsor (I
>>>> assume).
>>>>
>>>> Vladimir
>>>>
>>>> On 2/12/13 9:25 AM, Igor Ignatyev wrote:
>>>>> Do I need more review for this change or one is enough?
>>>>>
>>>>> Best regards,
>>>>> Igor Ignatyev
>>>>>
>>>>> On 02/07/2013 10:37 PM, Vladimir Kozlov wrote:
>>>>>> Nice.
>>>>>>
>>>>>> Thanks,
>>>>>> Vladimir
>>>>>>
>>>>>> On 2/7/13 10:36 AM, Igor Ignatyev wrote:
>>>>>>> Vladimir, thank you for review.
>>>>>>>
>>>>>>>  > We do not use 'mutable' in our sources. Why you need it?
>>>>>>>
>>>>>>> removed.
>>>>>>> forgot to remove it during my changes: at first i didn't remove
>>>>>>> 'const'
>>>>>>> from existing methods, so it was needed to use 'mutable'.
>>>>>>>
>>>>>>>  > In next code you can also use local fail_msg instead of
>>>>>>> overwriting
>>>>>>> _msg
>>>>>>>  > field:
>>>>>>>  >
>>>>>>>  > !     if (max_inline_size > default_max_inline_size) {
>>>>>>>  > !       set_msg("hot method too big");
>>>>>>>  >        }
>>>>>>>  > !     set_msg("too big");
>>>>>>>  > !     return false;
>>>>>>>
>>>>>>> fixed by moving 'set_msg("too big");' into else-branch
>>>>>>>
>>>>>>> http://cr.openjdk.java.net/~iignatyev/8007439/webrev.02/
>>>>>>>
>>>>>>> Best regards,
>>>>>>> Igor Ignatyev
>>>>>>>
>>>>>>> On 02/07/2013 10:11 PM, Vladimir Kozlov wrote:
>>>>>>>> This look good. Thanks!
>>>>>>>>
>>>>>>>> We do not use 'mutable' in our sources. Why you need it?
>>>>>>>>
>>>>>>>> In next code you can also use local fail_msg instead of overwriting
>>>>>>>> _msg
>>>>>>>> field:
>>>>>>>>
>>>>>>>> !     if (max_inline_size > default_max_inline_size) {
>>>>>>>> !       set_msg("hot method too big");
>>>>>>>>        }
>>>>>>>> !     set_msg("too big");
>>>>>>>> !     return false;
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Vladimir
>>>>>>>>
>>>>>>>> On 2/7/13 9:49 AM, Igor Ignatyev wrote:
>>>>>>>>> I have update the webrev.
>>>>>>>>>
>>>>>>>>>   - added _msg field  with getter and setter
>>>>>>>>>   - removed parameter 'msg' from InlineTree::print_inlining()
>>>>>>>>>   - methods InlineTree::should_inline(),
>>>>>>>>> InlineTree::should_not_inline()
>>>>>>>>> are no longer const-method
>>>>>>>>>
>>>>>>>>> http://cr.openjdk.java.net/~iignatyev/8007439/webrev.01/
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Best regards,
>>>>>>>>> Igor Ignatyev
>>>>>>>>>
>>>>>>>>> On 02/06/2013 06:20 AM, Vladimir Kozlov wrote:
>>>>>>>>>> Add _msg field into InlineTree class with getter and setter
>>>>>>>>>> instead of
>>>>>>>>>> passing is as parameter.
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Vladimir
>>>>>>>>>>
>>>>>>>>>> On 2/5/13 9:46 AM, Igor Ignatyev wrote:
>>>>>>>>>>> Hi all,
>>>>>>>>>>>
>>>>>>>>>>> Please review the patch.
>>>>>>>>>>>
>>>>>>>>>>> webrev: http://cr.openjdk.java.net/~iignatyev/8007439/webrev.00/
>>>>>>>>>>> testing: JPRT


More information about the hotspot-compiler-dev mailing list