Subzero. ARM32. Folding rematerializable offsets in address operands.

Subzero. ARM32. Folding rematerializable offsets in address operands.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4076
R=stichnot@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=866b6b19378691d0c07cd9c9b9c84036c201dfa2

[John, 2015-12-02 22:36:23]

Patchset #6 (id:100001) has been deleted

[John, 2015-12-02 22:37:35]

jpp@chromium.org changed reviewers:
+ kschimpf@google.com, sehr@chromium.org, stichnot@chromium.org

[John, 2015-12-02 22:37:36]

Once this lands I'll wrap up (hopefully) the sandbox work.

[Jim Stichnoth, 2015-12-03 16:44:33]

lgtm

https://codereview.chromium.org/1477873002/diff/120001/src/IceTargetLoweringA...
File src/IceTargetLoweringARM32.h (right):

https://codereview.chromium.org/1477873002/diff/120001/src/IceTargetLoweringA...
src/IceTargetLoweringARM32.h:837: // Ensures that the TempBase register held by
the this legalizer (if any)
Reflow to 80-col. :)

https://codereview.chromium.org/1477873002/diff/120001/tests_lit/assembler/ar...
File tests_lit/assembler/arm32/bic.ll (right):

https://codereview.chromium.org/1477873002/diff/120001/tests_lit/assembler/ar...
tests_lit/assembler/arm32/bic.ll:65: ; ASM-NEXT:  sub     r0, sp, #0
I realize there's probably no architectural difference, but I'm much happier
seeing an add of 0 instead of a subtract of 0 in this situation.  I'm not sure
how this particular diff came about, but would it be practical to change it?

[John, 2015-12-03 17:06:53]

https://codereview.chromium.org/1477873002/diff/120001/src/IceTargetLoweringA...
File src/IceTargetLoweringARM32.h (right):

https://codereview.chromium.org/1477873002/diff/120001/src/IceTargetLoweringA...
src/IceTargetLoweringARM32.h:837: // Ensures that the TempBase register held by
the this legalizer (if any)
On 2015/12/03 16:44:33, stichnot wrote:
> Reflow to 80-col. :)

This is /technically/ correct. :)
Done.

https://codereview.chromium.org/1477873002/diff/120001/tests_lit/assembler/ar...
File tests_lit/assembler/arm32/bic.ll (right):

https://codereview.chromium.org/1477873002/diff/120001/tests_lit/assembler/ar...
tests_lit/assembler/arm32/bic.ll:65: ; ASM-NEXT:  sub     r0, sp, #0
On 2015/12/03 16:44:33, stichnot wrote:
> I realize there's probably no architectural difference, but I'm much happier
> seeing an add of 0 instead of a subtract of 0 in this situation.  I'm not sure
> how this particular diff came about, but would it be practical to change it?

Done.

[John, 2015-12-03 17:45:35]

Description was changed from

==========
Subzero. ARM32. Folding rematerializable offsets in address operands.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4076
==========

to

==========
Subzero. ARM32. Folding rematerializable offsets in address operands.

BUG= https://code.google.com/p/nativeclient/issues/detail?id=4076
R=stichnot@chromium.org

Committed:
https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a...
==========

[John, 2015-12-03 17:45:37]

Committed patchset #8 (id:160001) manually as
866b6b19378691d0c07cd9c9b9c84036c201dfa2 (presubmit successful).