RFR(XS): 8246369: CodeCache.findBlobUnsafe(addr) sometimes asserts with valid address

[Chris Plummer]

Thanks!

On 6/16/20 4:40 PM, Alex Menkov wrote:
> Hi Chris,
>
> LGTM.
>
> --alex
>
> On 06/16/2020 14:10, Chris Plummer wrote:
>> Ping! Can I get one more review please? This is a very simple change 
>> and doesn't really require knowing much about SA. I bit of 
>> understanding of hotspot native heap memory allocations helps a bit 
>> (HeapBlock in heap.hpp).
>>
>> thanks,
>>
>> Chris
>>
>> On 6/15/20 4:40 PM, Chris Plummer wrote:
>>> I'll fix the indenting.
>>>
>>> Thanks for the review!
>>>
>>> Chris
>>>
>>> On 6/15/20 3:12 PM, serguei.spitsyn at oracle.com wrote:
>>>> Hi Chris,
>>>>
>>>> It looks good.
>>>>   134     if (Assert.ASSERTS_ENABLED) {
>>>> 135 // The pointer to the HeapBlock that contains this blob is 
>>>> outside of the blob,
>>>> 136 // but it shouldn't be an error to find a blob based on the 
>>>> pointer to the HeapBlock.
>>>> 137 // The heap block header is padded out to an 8-byte boundary. 
>>>> See heap.hpp. The
>>>> 138 // simplest way to compute the header size is just 2 * 
>>>> addressSize.
>>>> 139 Assert.that(result.blobContains(start) ||
>>>> 140 result.blobContains(start.addOffsetTo(2 * 
>>>> VM.getVM().getAddressSize())),
>>>>   141                     "found wrong CodeBlob");
>>>>   142     }
>>>> The lines 139-141 have wrong indent.
>>>> No need for another webrev.
>>>>
>>>> Thanks
>>>> Serguei
>>>>
>>>>
>>>> On 6/15/20 13:03, Chris Plummer wrote:
>>>>> Hello,
>>>>>
>>>>> Please help review the following simple fix:
>>>>>
>>>>> http://cr.openjdk.java.net/~cjplummer/8246369/webrev.00/index.html
>>>>> https://bugs.openjdk.java.net/browse/JDK-8246369
>>>>>
>>>>> Details are in the CR description.
>>>>>
>>>>> thanks,
>>>>>
>>>>> Chris
>>>>
>>>
>>