RFR: JDK-8331859 : [PPC64] Remove support for Power7 and older [v10]

[Martin Doerr]

On Thu, 15 May 2025 07:56:00 GMT, Suchismith Roy <sroy at openjdk.org> wrote:

>> JBS Issue: [JDK-8331859](https://bugs.openjdk.org/browse/JDK-8331859) 
>> Linux PPC64le requires Power8 since the beginning.
>> AIX requires Power8 with the new OpenXL based build ([JDK-8307520](https://bugs.openjdk.org/browse/JDK-8307520)). The old build has been removed in JDK 23 ([JDK-8327701](https://bugs.openjdk.org/browse/JDK-8327701)).
>> Linux PPC64 Big Endian is no longer officially supported (only kept alive for development, debugging and testing purposes).
>> 
>> The following checks for old processors are no longer needed:
>> 8: VM_Version::has_lqarx()
>> 7: VM_Version::has_popcntw()
>> 6: VM_Version::has_cmpb()
>> 5: VM_Version::has_popcntb()
>> These ones and some more checks for old instructions are no longer needed. All code which is no longer reachable when removing them should also get removed.
>> Checks like "PowerArchitecturePPC64 >= 8" (or older) can be removed.
>> 
>> Atomic::PlatformCmpxchg<1>::operator() can be simplified by using sub-word instructions (lharx, lbarx).
>> 
>> Temp registers can be removed from cmpxchgb and cmpxchgh.
>> 
>> Build flags "-mcpu=powerpc64 -mtune=power5" for Big Endian linux should get replaced by "-mcpu=power8 -mtune=power8" as already used for linux PPC64le.
>
> Suchismith Roy has updated the pull request incrementally with one additional commit since the last revision:
> 
>   adapt vsx

Very nice! Thanks! I only have some small requests remaining. We should also run tests on Power8 and with -XX:-SuperwordUseVSX.

src/hotspot/cpu/ppc/c1_LIRGenerator_ppc.cpp line 736:

> 734:         case vmIntrinsics::_dsqrt:
> 735:         case vmIntrinsics::_dsqrt_strict:
> 736:           /* run interpreted */

These 2 cases should be removed. They are already handled above, now.

src/hotspot/cpu/ppc/matcher_ppc.hpp line 161:

> 159: 
> 160:   // true means we have fast l2f conversion
> 161:   static bool convL2FSupported(void) {

Please make it `static constexpr` like on the other platforms.

src/hotspot/cpu/ppc/vm_version_ppc.cpp line 68:

> 66:       FLAG_SET_ERGO(PowerArchitecturePPC64, 9);
> 67:     } else {
> 68:       FLAG_SET_ERGO(PowerArchitecturePPC64, 0);

I think we should set it to 8, here.

src/hotspot/cpu/ppc/vm_version_ppc.cpp line 76:

> 74:     case 10: if (!VM_Version::has_brw()    ) break;
> 75:     case  9: if (!VM_Version::has_darn()   ) break;
> 76:     case  0: PowerArchitecturePPC64_ok = true; break;

And replace 0 by 8. 0 is no longer supported.

src/hotspot/cpu/ppc/vm_version_ppc.cpp line 115:

> 113:     }
> 114:     if (FLAG_IS_DEFAULT(UseVectorByteReverseInstructionsPPC64)) {
> 115:       FLAG_SET_ERGO(UseVectorByteReverseInstructionsPPC64, true);

I think this part should be reverted. It is still possible to switch off SuperwordUseVSX.

src/hotspot/cpu/ppc/vm_version_ppc.cpp line 253:

> 251:   }
> 252: 
> 253: 

Please remove the extra newline.

src/hotspot/cpu/ppc/vm_version_ppc.cpp line 274:

> 272: 
> 273:   if (FLAG_IS_DEFAULT(UseSHA256Intrinsics)) {
> 274:     FLAG_SET_DEFAULT(UseSHA256Intrinsics, true);

I think UseSHA256Intrinsics should only get enabled if UseSHA is enabled.

src/hotspot/cpu/ppc/vm_version_ppc.cpp line 278:

> 276: 
> 277:   if (FLAG_IS_DEFAULT(UseSHA512Intrinsics)) {
> 278:     FLAG_SET_DEFAULT(UseSHA512Intrinsics, true);

Same here.

src/hotspot/os_cpu/linux_ppc/atomic_linux_ppc.hpp line 257:

> 255:     "   lbarx   %[old_value], 0, %[dest]                  \n"
> 256:     /* extract byte and compare */
> 257:     "   cmpw    %[compare_value], %[old_value] \n"

Please increase indentation before \n.

src/hotspot/os_cpu/linux_ppc/atomic_linux_ppc.hpp line 266:

> 264:     /* out */
> 265:     : [old_value]           "=&r"   (old_value),
> 266:       [loaded_value]             "=&r"   (loaded_value),

Indentation.

-------------

PR Review: https://git.openjdk.org/jdk/pull/20262#pullrequestreview-2843573563
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091099298
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091107496
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091140960
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091144394
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091149721
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091152394
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091156968
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091158142
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091164357
PR Review Comment: https://git.openjdk.org/jdk/pull/20262#discussion_r2091165323