Seeing error in one test on ARM: org.openjdk.jcstress.tests.volatiles.ReadTwiceOverVolatileReadTest

[Stuart Monteith]

Hello,
  Would testing with -XX:-UseOnStackReplacement eliminate one of those
possibilities?
I regularly test with JCStress, but unfortunately not on ThunderX, and
with C2 only.

BR,

   Stuart


On 23 February 2017 at 10:19, Aleksey Shipilev <shade at redhat.com> wrote:
> Hi Derek,
>
> On 02/22/2017 11:49 PM, White, Derek wrote:
>> Could this be a test bug? Thanks!
>
> From the JMM standpoint, this is a test bug. JMM allows (1, 0, 0) there. It is
> also obvious we have a missing outcome, because there are 8 cases to cover, and
> we only have 7.
>
> Fixed:
>  http://hg.openjdk.java.net/code-tools/jcstress/rev/f4757f44c8b6
>
> What I find alarming though, is that generated code has reads in actor2() in
> order. Having a rare case when you see non-coherent read of $x must mean one of
> two things: a) there is an intermediate OSR stub that got compiled a little
> differently -- and this is fine; b) your host machine is not always coherent --
> and this is *not* fine for ARM.
>
> Coherency failures should have been caught by Opaque tests, like JMMSample_03:
>  http://hg.openjdk.java.net/code-tools/jcstress/file/f4757f44c8b6/jcstress-samples/src/main/java/org/openjdk/jcstress/samples/JMMSample_03_Coherence.java
>
> Thanks,
> -Aleksey
>
>> public class ReadTwiceOverVolatileReadTest {
>>     int x;
>>     volatile int y;
>>
>>     @Actor
>>     public void actor1() {
>>         x = 1;
>>         y = 1;
>>     }
>>     @Actor
>>     public void actor2(IntResult3 r) {
>>         r.r1 = x;
>>         r.r2 = y;
>>         r.r3 = x;
>>     }
>> }
>
> ...
>
>> === actor2() ===
>> --- C1 ---:
>>   0x0000ffff6c9ad068: ldr     w11, [x1,#12]   ;*getfield x {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 63)
>>   0x0000ffff6c9ad06c: cbz     x2, 0x0000ffff6c9ad09c
>>                                                 ;*putfield r1 {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 63)
>>   0x0000ffff6c9ad070: add     x10, x1, #0x10
>>   0x0000ffff6c9ad074: ldar    w10, [x10]
>>   0x0000ffff6c9ad078: str     w11, [x2,#12]   ;*getfield y {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 64)
>>   0x0000ffff6c9ad07c: ldr     w12, [x1,#12]
>>   0x0000ffff6c9ad080: str     w12, [x2,#20]   ;*putfield r3 {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 65)
>>
>>   0x0000ffff6c9ad084: str     w10, [x2,#16]   ;*getfield x {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 63)
>> --- C2 ---:
>>   0x0000ffff87d8923c: ldr     w0, [x1,#12]    ;*getfield x {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 63)
>>   0x0000ffff87d89240: str     w0, [x2,#12]    ;*putfield r1 {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 63)
>>                                                 ; implicit exception:
>>   0x0000ffff87d89244: ldr     w0, [x1,#16]
>>   0x0000ffff87d89248: dmb     ishld           ;*getfield y {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 64)
>>   0x0000ffff87d8924c: str     w0, [x2,#16]    ;*putfield r2 {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 64)
>>   0x0000ffff87d89250: ldr     w0, [x1,#12]    ;*getfield x {reexecute=0 rethrow=0 return_oop=0}
>>                                                 ; - (line 65)
>>   0x0000ffff87d89254: str     w0, [x2,#20]    ;*putfield r3 {reexecute=0 rethrow=0 return_oop=0}
>>
>