Backport hash collision workaround to 3.6.

Backport string hash collision workaround to 3.6.
This is made up of 9956, 10351, 10338, 10330 and 10362.
This change bakes the string hash key into the snapshot, so
it is determined at build time for shapshot configs.
Committed: http://code.google.com/p/v8/source/detail?r=10363

[Vyacheslav Egorov (Chromium), 2012-01-09 19:09:33]

lgtm

http://codereview.chromium.org/9124004/diff/1/src/arm/code-stubs-arm.cc
File src/arm/code-stubs-arm.cc (right):

http://codereview.chromium.org/9124004/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5398: __ and_(hash, hash,
Operand(kHashShiftCutOffMask));
Shouldn't it be SetCC? (instead of SetCC on the one above).

http://codereview.chromium.org/9124004/diff/1/src/mips/code-stubs-mips.cc
File src/mips/code-stubs-mips.cc (right):

http://codereview.chromium.org/9124004/diff/1/src/mips/code-stubs-mips.cc#new...
src/mips/code-stubs-mips.cc:5621: // if (hash == 0) hash = 27;
should not there be some special masking here?

http://codereview.chromium.org/9124004/diff/1/src/objects.h
File src/objects.h (right):

http://codereview.chromium.org/9124004/diff/1/src/objects.h#newcode6080
src/objects.h:6080: static const uint32_t kHashBitMask = 0xffffffffu >>
kHashShift;
this constant is duplicated in every code generator.

[Erik Corry, 2012-01-10 00:18:44]

http://codereview.chromium.org/9124004/diff/1/src/arm/code-stubs-arm.cc
File src/arm/code-stubs-arm.cc (right):

http://codereview.chromium.org/9124004/diff/1/src/arm/code-stubs-arm.cc#newco...
src/arm/code-stubs-arm.cc:5398: __ and_(hash, hash,
Operand(kHashShiftCutOffMask));
On 2012/01/09 19:09:33, Vyacheslav Egorov wrote:
> Shouldn't it be SetCC? (instead of SetCC on the one above).

Good catch, this needs fixing in bleeding edge too.

http://codereview.chromium.org/9124004/diff/1/src/mips/code-stubs-mips.cc
File src/mips/code-stubs-mips.cc (right):

http://codereview.chromium.org/9124004/diff/1/src/mips/code-stubs-mips.cc#new...
src/mips/code-stubs-mips.cc:5621: // if (hash == 0) hash = 27;
On 2012/01/09 19:09:33, Vyacheslav Egorov wrote:
> should not there be some special masking here?

As noted in the description I wasn't able to get MIPS working correctly, so
there may be a followup patch for MIPS (it was already broken in this branch if
I understand correctly, the external MIPS git repo is apparently the place to go
for 3.6).

http://codereview.chromium.org/9124004/diff/1/src/objects.h
File src/objects.h (right):

http://codereview.chromium.org/9124004/diff/1/src/objects.h#newcode6080
src/objects.h:6080: static const uint32_t kHashBitMask = 0xffffffffu >>
kHashShift;
On 2012/01/09 19:09:33, Vyacheslav Egorov wrote:
> this constant is duplicated in every code generator.

Yes, this should be fixed in bleeding edge.