Refactor the code cache to handle large number of properties on the global ob...

Refactor the code cache to handle large number of properties on the global object.

A separate object type for the code cache have been added. This object has two different code caches. The first one (default_cache) is a fixed array organized in the same way as the as the code cache was before. The second cache (global_access_cache) is for code stubs to access the global object. This cache is organized as a hash table taking the property name and code flags as the key.

The reason for separating the global access stubs into a hash table representation is that the number of these is not bounded in the same was as the other types.

BUG=613
Committed: http://code.google.com/p/v8/source/detail?r=3952

[Mads Ager (chromium), 2010-02-25 13:26:24]

LGTM

http://codereview.chromium.org/652119/diff/2001/3002
File src/objects.cc (right):

http://codereview.chromium.org/652119/diff/2001/3002#newcode3078
src/objects.cc:3078: if (code->type() == NORMAL) {
It is not completely trivial that MONOMORPHIC & NORMAL implies that the code is
for accessing a global property.  Could you add a comment just stating that this
is the case.  It is clear however, that MONOMORPHIC & NORMAL means that there
can be an unbounded number of them and we therefore need a hash table.

http://codereview.chromium.org/652119/diff/2001/3002#newcode3079
src/objects.cc:3079: // Make sure the a hash table is allocated got the global
access code cache.
the -> that
got -> for?

http://codereview.chromium.org/652119/diff/2001/3002#newcode3202
src/objects.cc:3202: // This is not used for global object IC's.
Should we just assert that code->type() != NORMAL in that case?

http://codereview.chromium.org/652119/diff/2001/3002#newcode3228
src/objects.cc:3228: // code object. The actual match is on the name and the
code flags. If this key
this -> a

http://codereview.chromium.org/652119/diff/2001/3002#newcode3229
src/objects.cc:3229: // is created using the flags in not a code object it can
only be used for
in -> and

http://codereview.chromium.org/652119/diff/2001/3003
File src/objects.h (right):

http://codereview.chromium.org/652119/diff/2001/3003#newcode3733
src/objects.h:3733: // Lookup Code object in the cache. Returns code object if
found.
Code -> code.

Document what it returns when not found?

http://codereview.chromium.org/652119/diff/2001/3003#newcode3737
src/objects.h:3737: // code object is not in that cache. This index can be ised
to later call
ised -> used

http://codereview.chromium.org/652119/diff/2001/3003#newcode3763
src/objects.h:3763: // Code cache layout. The code cache is a fixed array where
the first element
This comment looks wrong - left over from a previous version of this change?

[Søren Thygesen Gjesse, 2010-02-25 14:00:07]

http://codereview.chromium.org/652119/diff/2001/3002
File src/objects.cc (right):

http://codereview.chromium.org/652119/diff/2001/3002#newcode3078
src/objects.cc:3078: if (code->type() == NORMAL) {
On 2010/02/25 13:26:24, Mads Ager wrote:
> It is not completely trivial that MONOMORPHIC & NORMAL implies that the code
is
> for accessing a global property.  Could you add a comment just stating that
this
> is the case.  It is clear however, that MONOMORPHIC & NORMAL means that there
> can be an unbounded number of them and we therefore need a hash table.

Renamed GlobalAccess -> NormalType global_access -< normal_type, and added a
comment indicating that this is property cell access.

http://codereview.chromium.org/652119/diff/2001/3002#newcode3079
src/objects.cc:3079: // Make sure the a hash table is allocated got the global
access code cache.
On 2010/02/25 13:26:24, Mads Ager wrote:
> the -> that
> got -> for?

Done.

http://codereview.chromium.org/652119/diff/2001/3002#newcode3202
src/objects.cc:3202: // This is not used for global object IC's.
On 2010/02/25 13:26:24, Mads Ager wrote:
> Should we just assert that code->type() != NORMAL in that case?

Done.

http://codereview.chromium.org/652119/diff/2001/3002#newcode3228
src/objects.cc:3228: // code object. The actual match is on the name and the
code flags. If this key
On 2010/02/25 13:26:24, Mads Ager wrote:
> this -> a

Done.

http://codereview.chromium.org/652119/diff/2001/3002#newcode3229
src/objects.cc:3229: // is created using the flags in not a code object it can
only be used for
On 2010/02/25 13:26:24, Mads Ager wrote:
> in -> and

Done.

http://codereview.chromium.org/652119/diff/2001/3003
File src/objects.h (right):

http://codereview.chromium.org/652119/diff/2001/3003#newcode3733
src/objects.h:3733: // Lookup Code object in the cache. Returns code object if
found.
On 2010/02/25 13:26:24, Mads Ager wrote:
> Code -> code.
> 
> Document what it returns when not found?

Done.

http://codereview.chromium.org/652119/diff/2001/3003#newcode3737
src/objects.h:3737: // code object is not in that cache. This index can be ised
to later call
On 2010/02/25 13:26:24, Mads Ager wrote:
> ised -> used

Done.

http://codereview.chromium.org/652119/diff/2001/3003#newcode3763
src/objects.h:3763: // Code cache layout. The code cache is a fixed array where
the first element
On 2010/02/25 13:26:24, Mads Ager wrote:
> This comment looks wrong - left over from a previous version of this change?

Sorry about that, fixed.