This reduces chances of improper usage, see http://code.google.com/p/v8/issues/detail?id=586
for more details.

BUG=586

Review URL: http://codereview.chromium.org/555072

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3696 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

As an afterthought, I realized that I put function objects moves
reporting into a method that deals with only code object moves.  I've
looked up that function objects are allocated in old pointer space and
new space, so I moved logging to the corresponding VM methods.

BUG=553

Review URL: http://codereview.chromium.org/552089

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3679 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The problem appeared due to a fact that stubs doesn't create a stack
frame, reusing the stack frame of the caller function. When building
stack traces, the current function is retrieved from PC, and its
callees are retrieved by traversing the stack backwards. Thus, for
stubs, the stub itself was discovered via PC, and then stub's caller's
caller was retrieved from stack.

To fix this problem, a pointer to JSFunction object is now captured
from the topmost stack frame, and is saved into stack trace log
record. Then a simple heuristics is applied whether a referred
function should be added to decoded stack, or not, to avoid reporting
the same function twice (from PC and from the pointer.)

BUG=553
TEST=added to mjsunit/tools/tickprocessor

Review URL: http://codereview.chromium.org/546089

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3673 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Always invoke HeapObjectIterator::has_next() before invoking HeapObjectIterator::next().
This is necessary as ::has_next() has an important side-effect of going to the next
page when current page is exhausted.

And to find if pointers are encodable use more precise data---top of map space, not a number
of pages, as pages might stay in map space due to chunking.

Review URL: http://codereview.chromium.org/552066

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3672 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

drop below threshold.

Review URL: http://codereview.chromium.org/509035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3600 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Force mark sweep instead of compcation if size of map space is too big to allow forward pointers encoding.

Review URL: http://codereview.chromium.org/507025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3497 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

On 32-bit the maps are now aligned on a 32-byte boundary in order to encode more maps during compacting GC. The actual size of a map on 32-bit is 28 bytes making this change waste 4 bytes per map.

On 64-bit the encoding for compacting GC is now using more than 32-bits and the maps here are still pointer size aligned. The actual size of a map on 64-bit is 48 bytes and this change does not intruduce any waste.

My choice of 16 bits for kMapPageIndexBits for 64-bit should give the same maximum number of pages (8K) for map space. As maps on 64-bit are larger than on 32-bit the total number of maps on 64-bit will be smaller than on 32-bit. We could consider raising this to 17 or 18.

I moved the kPageSizeBits to globals.h as the calculation of the encoding really depended on this.

There are still an #ifdef/#endif in objects.h and this constant could be moved to globaks.h as well, but I kept it together with the related constants.

All the tests run in debug mode with additional options --gc-global --always-compact as well (except for a few tests on which also fails before this change when run with --gc-global --always-compact).

BUG=http://code.google.com/p/v8/issues/detail?id=524
BUG=http://crbug.com/29428
TEST=test/mjsunit/regress/regress-524.js

Review URL: http://codereview.chromium.org/504026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3481 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Instead of weak handles external strings use a separate table.  This
table uses 5 times less memory than weak handles.  Moreover, since we
don't have to follow the weak handle callback protocol we can collect
the strings faster and even on scavenge collections.

Review URL: http://codereview.chromium.org/467037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3439 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Set warning level to /W3 and change implicit conversions from size_t
to int. Most "fixes" are simply manifesting the implicit casts or using
a special strlen replacement that returns int.

Review URL: http://codereview.chromium.org/390004


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3273 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

As a first step to reduce the complexity of the string hierachy the sliced string type is removed. Whenever a sub-string is created it is allocated as a fresh flat string.
Review URL: http://codereview.chromium.org/385004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3259 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/355041


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3230 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix X64 build in the case that debugger support is disabled.  Change function name from IsCallInstruction to IsPatchedReturnSequence on all platforms.
Review URL: http://codereview.chromium.org/267116

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3072 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/200095

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3021 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Stop "cooking" targets of jumps and calls in code objects.  Do not convert jump and call targets to absolute pointers to Code objects during GC, heap verification, and serialization.
Review URL: http://codereview.chromium.org/203070

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2941 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/193111

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2888 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

in old space. Before this change we would compute the fragmentation 
limit before the GC, but that means that we take all the garbage 
into account - and consequently we almost never hit the limit (because 
we have a lot of garbage).

This change changes the policy to compact on the *next* GC if we
determine that we have too much fragmentation after doing a GC.
We'll do a GC if we've wasted more than 1MB and more than 15% of
the old space size.

This can be further improved by computing whether or not to compact
during the marking phase.
Review URL: http://codereview.chromium.org/202008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2841 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

initialized.

Minor cleanups.

Review URL: http://codereview.chromium.org/173465

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2757 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Android system provides a unique feature that it sends a notification to the
browser in low memory condition, and the browser cleans up cache and frees
resources. Forcing a GC in low memory condition can free DOM objects and also
can shrink the old spaces.

This patch addresses the last comment in 
http://codereview.chromium.org/173016/show

	
Mads Ager 2009/08/19 17:24:23 
I would prefer to not use the flags to signal that a compacting collection is
requested.

TBR = ager
Review URL: http://codereview.chromium.org/173102

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2725 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

space is similar to map space in that it has fixed-size objects.  A
common superclass for a space with fixed size objects is used for the
map space and cell space.

Allocate all cells in cell space.  Handle it during all GCs.  Modify
the free-list node representation (so that the size is not at a fixed
offset in all cells) to allow two-pointer free-list nodes.  Clean up
some stuff in the MC collector.

Review URL: http://codereview.chromium.org/155211

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2411 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

register on ARM.
* Make some compile-time loops into run-time loops for compactness.
Review URL: http://codereview.chromium.org/149324

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2398 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Remove references to Array:kHeaderSize.

Review URL: http://codereview.chromium.org/150098


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2303 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This avoids back-to-back mark-sweep collections.

Review URL: http://codereview.chromium.org/136001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2219 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This issue was raised by Brett Wilson while reviewing my changelist for readability. Craig Silverstein (one of C++ SG maintainers) confirmed that we should declare one namespace per line. Our way of namespaces closing seems not violating style guides (there is no clear agreement on it), so I left it intact.

Review URL: http://codereview.chromium.org/115756


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2038 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/113333

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1932 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TBR=ager@chromium.org

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1862 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

table debacle.

Review URL: http://codereview.chromium.org/109012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1856 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

are not treated as roots, but all their subparts are.
Review URL: http://codereview.chromium.org/108002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1849 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Mark all objects reachable from the symbols except the symbols
themselves as live (and reachable from strong roots).  This ensures
that if the symbol itself remains alive for any reason, and if it was
a sliced string or cons string backed by an external string, then the
external string will be strongly reachable and therefore not get a
weak reference callback.

Review URL: http://codereview.chromium.org/100344

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1846 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

resource to NULL when removing it from the symbol table.  This makes
sure that the debugger can recognize the external string as being
"deleted".  Now, whenever an external resource is deleted, the
resource pointer is set to NULL.

This is really a workaround of a debugger problem.  We need to make
sure that the debugger only finds scripts in the heap that are
actually live.
Review URL: http://codereview.chromium.org/69029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1734 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

GCs (not just mark-compacts) and make the mark-compact
shortcutting of cons-strings identical to the scavenge
version.
Review URL: http://codereview.chromium.org/67125

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1699 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

symbol table verification after mark-compact GCs.
Review URL: http://codereview.chromium.org/73029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1698 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

- Added special cutouts if a Vector has NULL data, which will now happen
  if an external string's resource has been deleted.
- Added an verification phase before old gen GC to verify that all real
  entries in the SymbolTable are valid symbols.
- Added test that verifies the correct behaviour of the workaround.

Review URL: http://codereview.chromium.org/66011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1689 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1589 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

- Pass the knowledge whether the old GC is compacting to the GC prologue and epilogue. This allows us to skip frame cooking and uncooking when doing a mark-sweep GC.
- Add the ability for the code to refer to its code object by adding a handle to the code object in the MacroAssembler.

Review URL: http://codereview.chromium.org/27133

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1368 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

create copies in the embedding code (aka WebKit V8 bindings) on every
external use.

Review URL: http://codereview.chromium.org/21117

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1252 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Renamed functions related to patching of code with call instructions to match the naming conversion.

BUG=1240753
Review URL: http://codereview.chromium.org/20176

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1239 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Added check for logging enabled in two places where processing/allocation was performed only to be used only as parameters for the LOG macro. Looked through all the uses of the LOG macro and found only these two places where additional checking was required.

Fixed a couple of comments looking through all the uses of the LOG macro.
Review URL: http://codereview.chromium.org/20018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1220 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

verification code and updating outdated comments.
Review URL: http://codereview.chromium.org/18502

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1131 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The main goal was to improve O(n^2) behavior when there are many object groups.  The old API required the grouping to be done on the v8 side, along with a linear search.  The new interface requires the caller to do the grouping, passing V8 entire groups at a time.  This removes the group id concept on the v8 side.

  - Changed AddObjectToGroup to AddObjectGroup.
  - Removed the group id concept from the V8 side.
  - Remove a static constructor while I'm here, lazily initialize
    the object groups list.
  - Cleaned up return by non-const references to return pointers.

Review URL: http://codereview.chromium.org/13341

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@965 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Moves the calls to tracer and UpdateLiveObjectCount inside the call to
(inlined) SetMark.  Removes global object counter.
Review URL: http://codereview.chromium.org/8910

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@688 ce2b1a6d-e550-0410-aec6-3dcde31c8c00