MIPS: Fix FP load/store with large offsets from base register.

MIPS: Fix FP load/store with large offsets from base register.

BUG=481519
LOG=y

Committed: https://crrev.com/a601987024aaaf13aebadb29ffba3613738a128a
Cr-Commit-Position: refs/heads/master@{#28113}

[paul.l..., 2015-04-27 23:38:50]

paul.lind@imgtec.com changed reviewers:
+ akos.palfi@imgtec.com, balazs.kilvady@imgtec.com,
dusan.milosavljevic@imgtec.com, jarin@chromium.org, machenbach@chromium.org

[paul.l..., 2015-04-27 23:38:50]

Google team: I have no idea if this should be logged or not. Let me know if I
should change it.

MIPS team: to cover boundary case in ldc1 & sdc1 with both pos & neg offsets, I
could not think of anything more clever than two is_int16 checks. Suggestions?

[Michael Achenbach, 2015-04-28 09:03:18]

Adding such issues to the change log is fine.

[dusmil.imgtec, 2015-04-28 10:31:13]

https://codereview.chromium.org/1111633002/diff/1/src/mips/assembler-mips.cc
File src/mips/assembler-mips.cc (right):

https://codereview.chromium.org/1111633002/diff/1/src/mips/assembler-mips.cc#...
src/mips/assembler-mips.cc:1998: if (is_int16(src.offset_) &&
is_int16(src.offset_ + kPointerSize)) {
No need for two is_int16 checks, later one is enough.
For the sake of correctness, it should be kIntSize.

[dusmil.imgtec, 2015-04-28 14:26:00]

On 2015/04/28 10:31:13, dusmil.imgtec wrote:
> https://codereview.chromium.org/1111633002/diff/1/src/mips/assembler-mips.cc
> File src/mips/assembler-mips.cc (right):
> 
>
https://codereview.chromium.org/1111633002/diff/1/src/mips/assembler-mips.cc#...
> src/mips/assembler-mips.cc:1998: if (is_int16(src.offset_) &&
> is_int16(src.offset_ + kPointerSize)) {
> No need for two is_int16 checks, later one is enough.
> For the sake of correctness, it should be kIntSize.

Correction, my suggestion cannot handle negative offsets.
I do not have better idea at the moment.

[paul.l..., 2015-04-28 14:28:09]

Change kPointerSize to kIntSize per suggestion. As noted, negative offsets need
the 2nd check.

[dusmil.imgtec, 2015-04-28 14:30:15]

LGTM.

[paul.l..., 2015-04-28 14:41:19]

The CQ bit was checked by paul.lind@imgtec.com

[commit-bot: I haz the power, 2015-04-28 14:41:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1111633002/20001

[commit-bot: I haz the power, 2015-04-28 15:42:31]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-04-28 15:42:43]

Patchset 2 (id:??) landed as
https://crrev.com/a601987024aaaf13aebadb29ffba3613738a128a
Cr-Commit-Position: refs/heads/master@{#28113}