Simplifies MIPS LoadWordFromPoolOffset

Simplifies MIPS LoadWordFromPoolOffset
Committed: https://code.google.com/p/dart/source/detail?r=20539

[zra, 2013-03-26 18:50:43]

Changes for comments after I submitted the last CL

Thanks!
Zach

[regis, 2013-03-26 19:10:56]

https://codereview.chromium.org/13094015/diff/1/runtime/vm/assembler_mips.cc
File runtime/vm/assembler_mips.cc (right):

https://codereview.chromium.org/13094015/diff/1/runtime/vm/assembler_mips.cc#...
runtime/vm/assembler_mips.cc:75: lw(rd, Address(rd));
I had suggested  lw(rd, Address(PP, offset_low)) instead of the last 3
instructions. Is that not equivalent?

[zra, 2013-03-26 20:00:21]

https://codereview.chromium.org/13094015/diff/1/runtime/vm/assembler_mips.cc
File runtime/vm/assembler_mips.cc (right):

https://codereview.chromium.org/13094015/diff/1/runtime/vm/assembler_mips.cc#...
runtime/vm/assembler_mips.cc:75: lw(rd, Address(rd));
On 2013/03/26 19:10:56, regis wrote:
> I had suggested  lw(rd, Address(PP, offset_low)) instead of the last 3
> instructions. Is that not equivalent?

Ack =( You're right, they are. Sorry. I was confused because of a problem I ran
into earlier when the low bits of the offset were unsigned. As a separate
question, is it legal for the offset to be negative?

[regis, 2013-03-26 20:13:42]

On 2013/03/26 20:00:21, zra wrote:
> https://codereview.chromium.org/13094015/diff/1/runtime/vm/assembler_mips.cc
> File runtime/vm/assembler_mips.cc (right):
> 
>
https://codereview.chromium.org/13094015/diff/1/runtime/vm/assembler_mips.cc#...
> runtime/vm/assembler_mips.cc:75: lw(rd, Address(rd));
> On 2013/03/26 19:10:56, regis wrote:
> > I had suggested  lw(rd, Address(PP, offset_low)) instead of the last 3
> > instructions. Is that not equivalent?
> 
> Ack =( You're right, they are. Sorry. I was confused because of a problem I
ran
> into earlier when the low bits of the offset were unsigned. As a separate
> question, is it legal for the offset to be negative?

In general, yes, but in this particular case, all entries in the object pool
will always be at positive indices.

[regis, 2013-03-26 20:14:44]

LGTM

[zra, 2013-03-26 20:27:34]

Committed patchset #2 manually as r20539 (presubmit successful).