RFR 8150388: Remove SPARC 32-bit support

George Triantafillou george.triantafillou at oracle.com
Fri Apr 7 19:21:24 UTC 2017


Hi Vladimir,

On 4/7/2017 1:34 PM, Vladimir Kozlov wrote:
> Hi George,
>
> I noticed that you did not changed code guarded by NOT_LP64() and 
> LP64_ONLY() macros.
>
> Also the code in src/os_cpu/linux_sparc/vm is not changed.
My understanding is that this issue only covers solaris_sparc.

-George
>
> Vladimir
>
> On 4/7/17 10:26 AM, Vladimir Kozlov wrote:
>> On 4/7/17 8:04 AM, George Triantafillou wrote:
>>> Thanks Harold, I've made the suggested changes.
>>>
>>> -George
>>>
>>> On 4/7/2017 10:34 AM, harold seigel wrote:
>>>> Hi George,
>>>>
>>>> Most of the changes look good.  A few comments:
>>>>
>>>> 1. In stubGenerator_sparc.cpp, lines 831-835 should not be deleted,
>>>> just the "&& defined(_LP64)"
>>>>
>>>>    831 #if defined(ASSERT) && defined(_LP64)
>>>>    832 __ signx(Rint, Rtmp);
>>>>    833 __ cmp(Rint, Rtmp);
>>>>    834 __ breakpoint_trap(Assembler::notEqual, Assembler::xcc);
>>>>    835 #endif
>>
>> I still see this code is removed. Did you updated webrev?
>>
>>>>
>>>> 2. In atomic_solaris_sparc.hpp: remove lines 245 and 372
>>>>
>>>> 3. In prefetch_solaris_sparc.inline.hpp, remove lines 30 and 63
>>
>> Yes. We don't build JVM with C1 only for 64 bit.
>>
>>>>
>>>> 4. In thread_solaris_sparc.hpp, remove line 67.
>>
>> Yes.
>>
>> Other compiler code changes seems fine.
>>
>> Thanks,
>> Vladimir
>>
>>>>
>>>> Thanks, Harold
>>>>
>>>> On 4/7/2017 9:43 AM, George Triantafillou wrote:
>>>>> Please review this fix to remove SPARC 32-bit support.  Support for
>>>>> solaris-sparc has been dropped from the list of supported platforms.
>>>>>
>>>>> JBS: https://bugs.openjdk.java.net/browse/JDK-8150388
>>>>> webrev:
>>>>> http://cr.openjdk.java.net/~gtriantafill/8150388-webrev/webrev/
>>>>> <http://cr.openjdk.java.net/%7Egtriantafill/8150388-webrev/webrev/>
>>>>>
>>>>> Built and tested on solaris-sparcv9-debug,solaris-x64-debug with the
>>>>> nsk.jvmti, nsk.jdwp, and nsk.jdi testlists.
>>>>>
>>>>> Thanks.
>>>>>
>>>>> -George
>>>>>
>>>>
>>>



More information about the hotspot-compiler-dev mailing list