Implement clearing of CompareICs.

Implement clearing of CompareICs.

This allows CompareICs to be cleared during garbage collection to avoid
cross-context garbage retention through maps stored in CompareIC stubs
for the KNOWN_OBJECTS state.

R=vegorov@chromium.org
BUG=v8:2102
Committed: https://code.google.com/p/v8/source/detail?r=11491

[Michael Starzinger, 2012-04-30 13:48:58]

Only for ia32, ports to x64 and ARM will follow shortly.

[Vyacheslav Egorov (Chromium), 2012-04-30 14:27:10]

ia32 lgtm

https://chromiumcodereview.appspot.com/10263008/diff/1/src/ic.cc
File src/ic.cc (right):

https://chromiumcodereview.appspot.com/10263008/diff/1/src/ic.cc#newcode415
src/ic.cc:415: if (target->compare_state() == UNINITIALIZED) return;
We can limit clearing to those states that can retain objects.

https://chromiumcodereview.appspot.com/10263008/diff/1/src/ic.cc#newcode2469
src/ic.cc:2469: Code* CompareIC::initialize_stub(Token::Value op) {
I guess it should be called GetUnitializedStub() or something. (naming convetion
limits foo_bar methods to getters).

[Michael Starzinger, 2012-05-02 11:07:23]

Ported to x64 and ARM. Addressed comments. PTAL.

https://chromiumcodereview.appspot.com/10263008/diff/1/src/ic.cc
File src/ic.cc (right):

https://chromiumcodereview.appspot.com/10263008/diff/1/src/ic.cc#newcode415
src/ic.cc:415: if (target->compare_state() == UNINITIALIZED) return;
On 2012/04/30 14:27:10, Vyacheslav Egorov wrote:
> We can limit clearing to those states that can retain objects.

Done.

https://chromiumcodereview.appspot.com/10263008/diff/1/src/ic.cc#newcode2469
src/ic.cc:2469: Code* CompareIC::initialize_stub(Token::Value op) {
On 2012/04/30 14:27:10, Vyacheslav Egorov wrote:
> I guess it should be called GetUnitializedStub() or something. (naming
convetion
> limits foo_bar methods to getters).

Done. Renamed to GetRawUninitialized() to be in sync with the handlified version
a few lines below. We might consider renaming the stub accessors for LoadICs and
StoreICs as well at some point.

[Vyacheslav Egorov (Chromium), 2012-05-02 11:29:48]

lgtm

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/arm/ic-arm.cc
File src/arm/ic-arm.cc (right):

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/arm/ic-arm.cc#...
src/arm/ic-arm.cc:1732: // This is patching a conditional "jump if/if-not smi"
site.
I think to align with code snippets below this should be jump if not smi/jump if
smi.

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/x64/ic-x64.cc
File src/x64/ic-x64.cc (right):

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/x64/ic-x64.cc#...
src/x64/ic-x64.cc:1770: // short jump-if-carry/not-carry at this position.
Please update the comment just like in ia32

[Michael Starzinger, 2012-05-03 10:56:21]

Addressed comments. Rebased. Landed.

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/arm/ic-arm.cc
File src/arm/ic-arm.cc (right):

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/arm/ic-arm.cc#...
src/arm/ic-arm.cc:1732: // This is patching a conditional "jump if/if-not smi"
site.
On 2012/05/02 11:29:48, Vyacheslav Egorov wrote:
> I think to align with code snippets below this should be jump if not smi/jump
if
> smi.

Done.

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/x64/ic-x64.cc
File src/x64/ic-x64.cc (right):

https://chromiumcodereview.appspot.com/10263008/diff/12001/src/x64/ic-x64.cc#...
src/x64/ic-x64.cc:1770: // short jump-if-carry/not-carry at this position.
On 2012/05/02 11:29:48, Vyacheslav Egorov wrote:
> Please update the comment just like in ia32

Done.