Backports from jdk9

[Edward Nevill]

Hi,

I intend to backport the following changes from jdk9 to jdk8u.

With these changes jdk8u can build itself.

All the best,
Ed.


changeset:   7674:ad6e224490d8
user:        enevill
date:        Wed Dec 23 12:21:32 2015 +0000
files:       src/cpu/aarch32/vm/templateTable_aarch32.cpp
description:
8146094: aarch32: exception handling in putfield corrupts dispatch register
Summary: Use rscratch2 instead of rdispatch
Reviewed-by: duke
Contributed-by: triple.yang at linaro.org


changeset:   7675:f4a5655ece8a
user:        enevill
date:        Tue Dec 29 16:18:51 2015 +0000
files:       src/cpu/aarch32/vm/templateTable_aarch32.cpp
description:
8146284: integer shifts not correctly masked
Summary: and shift quantity with 0x1f as required by hava spec
Reviewed-by: duke
Contributed-by: xiang.yuan at linaro.org


changeset:   7676:87065cdd6c2e
user:        enevill
date:        Tue Dec 29 16:25:46 2015 +0000
files:       src/cpu/aarch32/vm/macroAssembler_aarch32.cpp
description:
8146285: aarch32: NegativeArraySizeException not thrown correctly
Summary: Restore correct value of lr in newarray handler
Reviewed-by: duke
Contributed-by: triple.yang at linaro.org, snazarkin at azul.com


changeset:   7677:0de080095246
user:        enevill
date:        Wed Dec 30 17:16:35 2015 +0000
files:       src/cpu/aarch32/vm/assembler_aarch32.cpp
description:
8146339: aarch32: mov immediate generates incorrect value for some immediate values
Summary: Use movw/movt instead of just movt for immediates with bottom 16 bits 0
Reviewed-by: duke
Contributed-by: xiang.yuan at linaro.org

changeset:   7678:2f30ec8989e9
user:        enevill
date:        Tue Jan 05 09:17:02 2016 +0000
files:       src/cpu/aarch32/vm/templateTable_aarch32.cpp
description:
8146456: aarch32: JVM crashes when run with -Xdebug option
Summary: Fix register corruption
Reviewed-by: duke
Contributed-by: shanyao.chen at linaro.org


changeset:   7679:f4b0c5814c62
tag:         tip
user:        jjoyce
date:        Sat Dec 26 22:30:15 2015 +0000
files:       src/cpu/aarch32/vm/assembler_aarch32.hpp
description:
8146193: Bad marshalling of single precision floats
Summary: Switched action of fp_rencode_real for 32/64 bit floats.
Reviewed-by: enevill