Remove PropertyCell space

Remove PropertyCell space
Replaces StoreGlobalCell / LoadGlobalCell with NamedField variants that use write barriers.
BUG=

Committed: https://crrev.com/16c8485a35efc36cf6ccd15185282d65412e1502
Cr-Commit-Position: refs/heads/master@{#27269}

[Toon Verwaest, 2015-03-17 16:43:56]

verwaest@chromium.org changed reviewers:
+ hpayer@chromium.org

[Toon Verwaest, 2015-03-17 16:43:57]

PTAL

[dcarney, 2015-03-17 19:22:56]

dcarney@chromium.org changed reviewers:
+ dcarney@chromium.org

[dcarney, 2015-03-17 19:22:57]

https://codereview.chromium.org/1016803002/diff/40001/src/objects-inl.h
File src/objects-inl.h (right):

https://codereview.chromium.org/1016803002/diff/40001/src/objects-inl.h#newco...
src/objects-inl.h:1898: DCHECK(!val->IsPropertyCell() && !val->IsCell());
the overloaded setter for PropertyCell seems dangerous
i think you want a DCHECK(!this->IsPropertyCell()) here

[yangguo, 2015-03-17 19:58:06]

yangguo@google.com changed reviewers:
+ yangguo@google.com

[yangguo, 2015-03-17 19:58:07]

There are a couple of comments in serialize.h that say that there is one opcode
for each space, of which there are 8. Might want to update them.

[Toon Verwaest, 2015-03-17 21:17:27]

Woops. I had updated that locally but forgot. Will do so.
About separating cell / PropertyCell, as discussed I'll look into this in a
separate CL.

[Hannes Payer (out of office), 2015-03-18 10:13:51]

LGTM

[Toon Verwaest, 2015-03-18 10:23:41]

The CQ bit was checked by verwaest@chromium.org

[Toon Verwaest, 2015-03-18 10:23:41]

The patchset sent to the CQ was uploaded after l-g-t-m from hpayer@chromium.org
Link to the patchset: https://codereview.chromium.org/1016803002/#ps100001
(title: " ")

[commit-bot: I haz the power, 2015-03-18 10:23:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1016803002/100001

[commit-bot: I haz the power, 2015-03-18 11:43:58]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2015-03-18 11:44:07]

Patchset 6 (id:??) landed as
https://crrev.com/16c8485a35efc36cf6ccd15185282d65412e1502
Cr-Commit-Position: refs/heads/master@{#27269}

[Benedikt Meurer, 2015-03-19 07:06:37]

Where's the design doc for this change? What are the implications?

[Toon Verwaest, 2015-03-19 12:56:56]

As far as I can tell this only affects a few microbenchmarks in sunspider, which
are buggy. They are buggy in the sense that they want to use local variables for
computation, but "forget" to put var in front of the variable name. E.g., "var
a1 = a2 = a3 = ..." makes a1 local, but a2, a3, ... global.

It affects them since now we need a write barrier. However, in all those
micro-benchmarks we store numbers. Representation tracking for globals will
allow us to avoid both the write barrier and the double allocation in
Crankshaft, hopefully restoring (and perhaps even improving beyond original)
performance.

The reason for removal was that it's just an extra source of fragmentation, and
causes extra GC overhead (splay actually improved slightly when I removed it).
The only cost is the write barrier, but "proper code" shouldn't write to globals
so frequently anyway. (It isn't measurable outside of sunspider.)

The reason why I disinherited PropertyCell from Cell is that they have nothing
to do with each other, apart from the fact that they both contain a value-field.
Both used to omit write-barriers, but now PropertyCell doesn't anymore. Putting
them in the same hierarchy became dangerous, as Dan noted in his comment on this
CL.