[15u] RFR 8255065: Zero: accessor_entry misses the IRIW case

[Doerr, Martin]

Hi Aleksey,

looks good. Thanks for fixing it.

Best regards,
Martin

> -----Original Message-----
> From: jdk-updates-dev <jdk-updates-dev-retn at openjdk.java.net> On
> Behalf Of Aleksey Shipilev
> Sent: Freitag, 23. Oktober 2020 13:14
> To: jdk-updates-dev at openjdk.java.net
> Subject: [15u] RFR 8255065: Zero: accessor_entry misses the IRIW case
> 
> Original bug:
>    https://bugs.openjdk.java.net/browse/JDK-8255065
>    https://git.openjdk.java.net/jdk/commit/bd45191f
> 
> Since cppInterpreter_zero.cpp was renamed to zeroInterpreter_zero.cpp in
> 16, this patch does not
> apply cleanly. So I had to apply the same hunk to the old location.
> 
> 15u patch:
> 
> diff -r 96851120ae5c src/hotspot/cpu/zero/cppInterpreter_zero.cpp
> --- a/src/hotspot/cpu/zero/cppInterpreter_zero.cpp      Thu Oct 08 08:10:55
> 2020 +0000
> +++ b/src/hotspot/cpu/zero/cppInterpreter_zero.cpp      Fri Oct 23 13:10:38
> 2020 +0200
> @@ -555,10 +555,13 @@
>       stack->overflow_check(1, CHECK_0);
>       stack->alloc(wordSize);
>       break;
>     }
>     if (entry->is_volatile()) {
> +    if (support_IRIW_for_not_multiple_copy_atomic_cpu) {
> +      OrderAccess::fence();
> +    }
>       switch (entry->flag_state()) {
>       case ctos:
>         SET_LOCALS_INT(object->char_field_acquire(entry->f2_as_index()), 0);
>         break;
> 
> Testing: Linux x86_64 Zero build
> 
> --
> Thanks,
> -Aleksey