Improve implementation of HSeqStringSetChar.

Improve implementation of HSeqStringSetChar.

This improves the generated code for HSeqStringSetChar across
all platforms, taking advantage of constant operands whenever
possible. It also drops the unused DefineSameAsFirst constraint
for the register allocator on x64 and ia32, where it caused
unnecessary spills when the string operand was live across the
HSeqStringSetChar instruction.

A new GVN flag StringChars is introduced to express dependencies
between HSeqStringSetChar, HStringCharCodeAt and the upcoming
HSeqStringGetChar (the GVNFlags type is now 64bit in size).

Also improves the test case.

TEST=mjsunit/string-natives
R=mstarzinger@chromium.org, yangguo@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=17521

[Benedikt Meurer, 2013-11-04 14:33:07]

Hey Yang,
This improves HSeqStringSetChar so we can actually use it for the hydrogen
StringAddStub. PTAL
-- Benedikt

[Benedikt Meurer, 2013-11-05 09:59:02]

Michi,
PTAL at the GVNFlags changes.
-- Benedikt

[Michael Starzinger, 2013-11-05 14:44:19]

LGTM on the GVN flag if there is no noticable performance impact due to the
increase in zone memory consumption.

[Yang, 2013-11-06 09:55:09]

lgtm if comments are addressed.

https://codereview.chromium.org/57383004/diff/110001/src/ia32/lithium-ia32.cc
File src/ia32/lithium-ia32.cc (right):

https://codereview.chromium.org/57383004/diff/110001/src/ia32/lithium-ia32.cc...
src/ia32/lithium-ia32.cc:1880: ? UseFixedOrConstant(instr->value(), eax)
Why does this have to be fixed to eax, why not UseRegisterOrConstant?

https://codereview.chromium.org/57383004/diff/110001/test/mjsunit/string-nati...
File test/mjsunit/string-natives.js (right):

https://codereview.chromium.org/57383004/diff/110001/test/mjsunit/string-nati...
test/mjsunit/string-natives.js:45: assertEquals("AAAAAAAAAAAAAAAAAAAAAAAAAB",
s2);
Please also add a test that tests non-latin1 characters and where the character
being set is not a constant.

[Benedikt Meurer, 2013-11-06 10:06:36]

https://codereview.chromium.org/57383004/diff/110001/src/ia32/lithium-ia32.cc
File src/ia32/lithium-ia32.cc (right):

https://codereview.chromium.org/57383004/diff/110001/src/ia32/lithium-ia32.cc...
src/ia32/lithium-ia32.cc:1880: ? UseFixedOrConstant(instr->value(), eax)
On 2013/11/06 09:55:09, Yang wrote:
> Why does this have to be fixed to eax, why not UseRegisterOrConstant?

Because mov_b() requires a byte register.

https://codereview.chromium.org/57383004/diff/110001/test/mjsunit/string-nati...
File test/mjsunit/string-natives.js (right):

https://codereview.chromium.org/57383004/diff/110001/test/mjsunit/string-nati...
test/mjsunit/string-natives.js:45: assertEquals("AAAAAAAAAAAAAAAAAAAAAAAAAB",
s2);
On 2013/11/06 09:55:09, Yang wrote:
> Please also add a test that tests non-latin1 characters and where the
character
> being set is not a constant.

That's done below with s3 already.

[Yang, 2013-11-06 10:24:59]

On 2013/11/06 10:06:36, Benedikt Meurer wrote:
> https://codereview.chromium.org/57383004/diff/110001/src/ia32/lithium-ia32.cc
> File src/ia32/lithium-ia32.cc (right):
> 
>
https://codereview.chromium.org/57383004/diff/110001/src/ia32/lithium-ia32.cc...
> src/ia32/lithium-ia32.cc:1880: ? UseFixedOrConstant(instr->value(), eax)
> On 2013/11/06 09:55:09, Yang wrote:
> > Why does this have to be fixed to eax, why not UseRegisterOrConstant?
> 
> Because mov_b() requires a byte register.
> 
>
https://codereview.chromium.org/57383004/diff/110001/test/mjsunit/string-nati...
> File test/mjsunit/string-natives.js (right):
> 
>
https://codereview.chromium.org/57383004/diff/110001/test/mjsunit/string-nati...
> test/mjsunit/string-natives.js:45: assertEquals("AAAAAAAAAAAAAAAAAAAAAAAAAB",
> s2);
> On 2013/11/06 09:55:09, Yang wrote:
> > Please also add a test that tests non-latin1 characters and where the
> character
> > being set is not a constant.
> 
> That's done below with s3 already.

lgtm.

[Benedikt Meurer, 2013-11-06 13:09:39]

Committed patchset #2 manually as r17521 (presubmit successful).