RFR: Fix, improve and refactor matrix barrier on AAarch64

[Roman Kennke]

Am 14.09.2017 um 16:42 schrieb Aleksey Shipilev:
> On 09/14/2017 04:39 PM, Roman Kennke wrote:
>> Am 14.09.2017 um 15:52 schrieb Aleksey Shipilev:
>>> *) Passing third parameter makes the method consistent with what? shenandoah_write_barrier_post in
>>> x86 uses rscratch1 without having it passed via register. g1_write_barrier_post also does not pass
>>> rscratch1, but uses it.
>> Most with itself. It seemed wrong to me to declare two temp regs as args, then use 3, and use 1
>> implicitely. Also, it made experimenting/changing temp reg while debugging easier. Do you want me to
>> change it back?
> Yeah, change it back: I'd like to be consistent across arches and collectors.
>
>> http://cr.openjdk.java.net/~rkennke/fix-matrix-barrier-aarch64/webrev.01/
> Good, except for rscratch1 change.

Ok, I made it a compromise: kept the method signature the same (i.e. 2 
tmp args), and declare tmp3 at the beginning of the code block to be 
rscratch1.

http://cr.openjdk.java.net/~rkennke/fix-matrix-barrier-aarch64/webrev.02/ 
<http://cr.openjdk.java.net/%7Erkennke/fix-matrix-barrier-aarch64/webrev.02/>

Good now?