RFR(s): 8221408: Windows 32bit build build errors/warnings in hotspot
David Holmes
david.holmes at oracle.com
Fri Mar 29 01:49:57 UTC 2019
+1
Thanks,
David
On 29/03/2019 9:58 am, Kim Barrett wrote:
>> On Mar 28, 2019, at 4:08 AM, Thomas Stüfe <thomas.stuefe at gmail.com> wrote:
>>
>> Hi Kim, David,
>>
>> new version:
>>
>> http://cr.openjdk.java.net/~stuefe/webrevs/8221408-win32-hotspot-buildfixes/webrev.01/webrev/
>>
>> Worked in feedback:
>>
>> - using PRAGMA_ macros instead of raw pragmas on os::current_stack_pointer()
>> - redefined the offending enum values in vmStruct.hpp to be static constants, for all platforms.
>>
>> Good?
>>
>> Did build successfully on 32bit windows. Submit tests are running but I do not expect any surprises.
>>
>> Thanks, Thomas
>
> ------------------------------------------------------------------------------
> src/hotspot/share/oops/markOop.hpp
> 148 const static uintptr_t hash_mask_in_place =
> 149 (address_word)hash_mask << hash_shift;
>
> The cast to address_word isn't needed. That was true previously too,
> but kept this copy consistent with that in the enum. That consistency
> is no longer a concern.
>
> Also, I'd like hash_mask and hash_mask_in_place moved up to after the
> enum with the other masks, rather than being separated from them by
> the intervening enum for biased_lock_alignment.
>
> ------------------------------------------------------------------------------
>
> Other than that, looks good. I don't need another webrev for those
> changes.
>
More information about the hotspot-runtime-dev
mailing list