RFR(XS) 8008750 [partfait] Null pointer deference in hotspot/src/share/vm/oops/instanceKlass.hpp

[Coleen Phillimore]

Oh, I see.  Can you still use assert rather then ShouldNotReachHere()?  
The former will vanish in product but the latter is added in product.

Coleen

On 03/05/2013 05:17 PM, Vladimir Kozlov wrote:
> We want to fix cases where Parfait gives "false" positives. That is 
> why we need real != NULL check, assert does not help.
>
> Vladimir
>
> On 3/5/13 1:50 PM, Coleen Phillimore wrote:
>> I can't see the parfait link in the bug.
>>
>> What does this change do but check for null in the cases where host
>> class and adr_implementers can't be null?
>>
>> In instanceKlass.hpp ShouldNotReachHere() is enabled in product so these
>> should be
>> assert(host != NULL, "not null");
>> *adr = host;
>>
>> I don't understand the motivation for this change and would rather not
>> enable these ShouldNotReachHere() in product mode, ie use asserts 
>> instead.
>>
>> Coleen
>>
>> On 03/05/2013 03:51 PM, Vladimir Kozlov wrote:
>>> Looks good. CC to runtime group.
>>>
>>> Vladimir
>>>
>>> On 3/5/13 12:37 PM, Morris Meyer wrote:
>>>> Folks,
>>>>
>>>> Could I get a review for this parfait issue? This has been through 
>>>> JPRT.
>>>>
>>>> Thanks much,
>>>>
>>>>          --mm
>>>>
>>>> WEBREV - http://cr.openjdk.java.net/~morris/8008750.01
>>>> BUG - https://jbs.oracle.com/bugs/browse/JDK-8008750
>>