MIPS: Fix Assembler::dd(Label*) implementation.

MIPS: Fix Assembler::dd(Label*) implementation.

BUG=
TEST=mjsunit/asm/embenchen/lua_binarytrees

Committed: https://crrev.com/3f7b797c1e066cd2a86508a051599624134f597a
Cr-Commit-Position: refs/heads/master@{#28303}

[balazs.kilvady, 2015-05-07 15:03:36]

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

[dusmil, 2015-05-07 15:31:02]

Can we use BlockGrowBufferScope?

[paul.l..., 2015-05-07 15:41:11]

On 2015/05/07 15:31:02, dusmil wrote:
> Can we use BlockGrowBufferScope?

I don't think you can block it, you might be emitting a large jump table, and
need to grow the buffer. How does this change work in that case?

[balazs.kilvady, 2015-05-07 16:23:45]

On 2015/05/07 15:41:11, paul.l... wrote:
> On 2015/05/07 15:31:02, dusmil wrote:
> > Can we use BlockGrowBufferScope?
> 
> I don't think you can block it, you might be emitting a large jump table, and
> need to grow the buffer. How does this change work in that case?

Assembler::dd(Label*) has a CheckBuffer() call at the first line which
guarantees that there is enough space in the buffer for an instruction
(buffer_space() >= kGap or call GrowBuffer() _before_ the new address computed
and inserted. Anyway if we don't need an emit()->?GrowBuffer() call in the if ()
branch then we don't need it in the else branch either. With huge jump tables
the dd(label) is called many times and it starts with a CheckBuffer() call at
each times so it must be OK I think.

[paul.l..., 2015-05-07 16:50:31]

paul.lind@imgtec.com changed reviewers:
+ danno@chromium.org

[paul.l..., 2015-05-07 16:50:32]

lgtm

[balazs.kilvady, 2015-05-07 16:56:22]

The CQ bit was checked by balazs.kilvady@imgtec.com

[commit-bot: I haz the power, 2015-05-07 16:56:40]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1127393003/1

[commit-bot: I haz the power, 2015-05-07 17:21:29]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-05-07 17:21:44]

Patchset 1 (id:??) landed as
https://crrev.com/3f7b797c1e066cd2a86508a051599624134f597a
Cr-Commit-Position: refs/heads/master@{#28303}