RFR(S): 8016277: Crash in nmethod::is_compiled_by_c1() on x86
Vladimir Kozlov
vladimir.kozlov at oracle.com
Wed Aug 21 15:16:58 PDT 2013
On 8/21/13 1:49 PM, Roland Westrelin wrote:
>> You did change the behavior of these methods - they don't return false now if _method == NULL.
>> Do you know why we rerun false before when _method == NULL?
>
> I assumed it was there to avoid the call to is_native_method() and a crash. Now that the is_native_method() call is gone, no need to check _method for NULL anymore. If the methods do what their names say they do then I don't see why we would care whether _method == NULL.
Then we don't need asserts now.
Thanks,
Vladimir
> That was part of "6912062: disassembler plugin needs to produce symbolic information in product mode" that John pushed.
>
> Roland.
>
More information about the hotspot-compiler-dev
mailing list