RFR: 8224162: assert(profile.count() == 0) failed: sanity in InlineTree::is_not_reached

[Jie Fu]

On 2019/5/22 上午11:57, Leonid Mesnik wrote:
> Could you please add reproducer as regression test for fix. Please verify that test failed without fix and pass with fix.

Done.

Updated: http://cr.openjdk.java.net/~jiefu/8224162/webrev.03/

# Testing
On an i7-8700 at 3.20GHz machine, the test results:
-----------------------------------------------
- fastdebug: make test 
TEST="test/hotspot/jtreg/compiler/profiling/TestProfileCounterOverflow.java" 
CONF=fastdebug
   Before fix: failed, elapsed time: 37s
   After  fix: pass,   elapsed time: 43s
- slowdebug: make test 
TEST="test/hotspot/jtreg/compiler/profiling/TestProfileCounterOverflow.java" 
CONF=slowdebug
   Before fix: failed, elapsed time: 38s
   After  fix: pass,   elapsed time: 43s
-----------------------------------------------

Please review it and give me some advice.

Thanks.
Best regards,
Jie

>
> Leonid
>
>> On May 21, 2019, at 8:54 PM, Jie Fu <fujie at loongson.cn> wrote:
>>
>> On 2019/5/21 下午10:30, Vladimir Ivanov wrote:
>>> It doesn't explain the failure since the assert is hit while parsing invoke* bytecode while type check failures are recorded for checkcast, aastore, and instanceof. Am I missing something important here?
>>>
>> Good question. I'm so sorry to make you confused.
>>
>> After a long time of digging into the code, I think this failure was cause by the overflow of profile.cout.
>>
>> A reproducer is constructed here:
>>   - http://cr.openjdk.java.net/~jiefu/8224162/CounterOverflow.java
>>
>> And I've changed the comment for the patch:
>>   - http://cr.openjdk.java.net/~jiefu/8224162/webrev.02/
>>
>> Please review it and give me some advice.
>>
>> Thanks.
>> Best regards,
>> Jie
>>
>>