[aarch64-port-dev ] Fwd: RFR(xs): 8233019: java.lang.Class.isPrimitive() (C1) returns wrong result if Klass* is aligned to 32bit

[Andrew Haley]

On 11/4/19 3:21 PM, Thomas Stüfe wrote:
> could some aarch64 people please take a quick look at this small patch?
> 
> The aarch64 part is really tiny, but I have no possibility to test this.
> 
> Last webrev:
> http://cr.openjdk.java.net/~stuefe/webrevs/8233019--c1-intrinsic-for-java.lang.class.isprimitive()-does-32bit-compare/webrev.01/webrev/
> Issue: https://bugs.openjdk.java.net/browse/JDK-8233019

Seems fine.

Before:

 ;;  block B0 [0, 4]
  0x0000ffffa1d93f54:   ldr	x0, [x1, #80]               ; implicit exception: dispatches to 0x0000ffffa1d93f78
  0x0000ffffa1d93f58:   cmp	w0, #0x0
  0x0000ffffa1d93f5c:   cset	x0, eq  // eq = none        ;*invokevirtual isPrimitive {reexecute=0 rethrow=0 return_oop=0}
                                                            ; - IsPrimitiveTest::isPrimitive at 1 (line 4)

After:

 ;;  block B0 [0, 4]
  0x0000ffff71dc75d4:   ldr	x0, [x1, #80]               ; implicit exception: dispatches to 0x0000ffff71dc75f8
  0x0000ffff71dc75d8:   cmp	x0, #0x0
  0x0000ffff71dc75dc:   cset	x0, eq  // eq = none        ;*invokevirtual isPrimitive {reexecute=0 rethrow=0 return_oop=0}
                                                           ; - IsPrimitiveTest::isPrimitive at 1 (line 4)

i.e. the first test is "cmp w0, #0x0", the second is "cmp x0, #0x0".
The first is a 32-bit comparison, the second 64-bit.

-- 
Andrew Haley  (he/him)
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
https://keybase.io/andrewhaley
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671