Commits · 03419c83696c2352791e44f3d3afc26360fbfb03 · Linshizhi / V8

25 Mar, 2009 10 commits

Simplify the way the code generator handles calls to IC stubs. Before · 03419c83

kmillikin@chromium.org authored Mar 25, 2009

we dispatched on the IC stub kind in a generic CallCodeObject
function.  Now, we have special functions for the load and store IC
stubs.

We also (for the load and store ICs) handle moving register arguments
into place only after the stack is prepared for the call.  This
replaces some memory-to-memory moves (for copies whose backing store
is occupied by a register needed for the arguments) with
memory-to-register moves.
Review URL: http://codereview.chromium.org/42602

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1613 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

03419c83

Add another check in profiler's stack iterator to ensure stability. · 49af1766

mikhail.naganov@gmail.com authored Mar 25, 2009

I made a test run of benchmarks and found another weak point. After
fixing it, benchmarks are able to run for a 100 times in a row
successfully in both builds.

Review URL: http://codereview.chromium.org/53050

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1612 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

49af1766

Added more checks to SafeStackFrameIterator to prevent crashes when profiling. · 9f5de7d9

mikhail.naganov@gmail.com authored Mar 25, 2009

Tested by profiling 3d-morph.js a 100 times both in debug and release builds.

Review URL: http://codereview.chromium.org/42600

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1611 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

9f5de7d9

Move reference to variable that isn't used on ARM. · eb3bf208

lrn@chromium.org authored Mar 25, 2009


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1610 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

eb3bf208

Introduce splaytree.KeyNotFoundError and use it for reporting · fc0af92f

kasperl@chromium.org authored Mar 25, 2009

issues when removing non-existing nodes from a SplayTree.
Review URL: http://codereview.chromium.org/42599

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1609 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

fc0af92f

Moved String.prototype.match implementation to C++. · bd8816ef

lrn@chromium.org authored Mar 25, 2009

Some extra runtime assertions added.


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1608 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

bd8816ef

The code generator doesn't need to put the argument count in eax for · 7c35a5d8

kmillikin@chromium.org authored Mar 25, 2009

call ICs. Refactor CallRuntime on ARM to match the structure of IA32.
Update some comments.
Review URL: http://codereview.chromium.org/42598

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1607 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

7c35a5d8

Make sure to generate a CodeCreateEvent for the CPU features · 4d8ea714

kasperl@chromium.org authored Mar 25, 2009

probe code object, and really raise an exception if SplayTree.Remove
is called with a key that cannot be found.
Review URL: http://codereview.chromium.org/42597

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1606 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

4d8ea714

Remove register counts from VirtualFrame, use register indices instead. · 36853c83

whesse@chromium.org authored Mar 25, 2009

Review URL: http://codereview.chromium.org/42565

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1605 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

36853c83

Use OpenThread instead of DuplicateHandle in Win32 profiler implementation. · 68d9782c

mikhail.naganov@gmail.com authored Mar 25, 2009

OpenThread doesn't fail in Chrome sandbox, while DuplicateHandle does.

Review URL: http://codereview.chromium.org/49028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1604 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

68d9782c

24 Mar, 2009 20 commits

Preparing push of version 1.1.3 to trunk. · 898297b0

kasperl@chromium.org authored Mar 24, 2009

Review URL: http://codereview.chromium.org/53012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1602 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

898297b0

Strcmp thinko and SConstruct improvement for d8 on Android. · 62a7f913

erik.corry@gmail.com authored Mar 24, 2009

Review URL: http://codereview.chromium.org/42560

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1601 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

62a7f913

Make the "has function literal" default value for loops conservative. · f38dc917

kmillikin@chromium.org authored Mar 24, 2009

It defaults to true so that if the analysis is not run we get a safe
analysis result.

Review URL: http://codereview.chromium.org/42562

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1600 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

f38dc917

Remove stl dependencies from d8. · 944258f9

erik.corry@gmail.com authored Mar 24, 2009

Review URL: http://codereview.chromium.org/42559

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1599 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

944258f9

Fix ARM simulator. Unless building for the actual ARM hardware, · 89a43b8f

kmillikin@chromium.org authored Mar 24, 2009

inline assembler in C++ code should be IA32 assembly, not ARM.

TBR=erik.corry@gmail.com

Review URL: http://codereview.chromium.org/53011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1597 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

89a43b8f

Fix ARM build for changes to VirtualFrame · 1d5f94c7

whesse@chromium.org authored Mar 24, 2009

Review URL: http://codereview.chromium.org/53009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1596 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

1d5f94c7

Add basic infrastructure for protecting V8's heap when leaving the VM · bc3fb118

kmillikin@chromium.org authored Mar 24, 2009

and unprotecting it when (re)entering. The functionality is enabled
by the flag --protect-heap and requires V8 to be built with
ENABLE_HEAP_PROTECTION and ENABLE_LOGGING_AND_PROFILING defined.

Implemented on Linux and Windows but not yet for other platforms.

Review URL: http://codereview.chromium.org/53004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1595 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

bc3fb118

Fix issue 286. Ensure frame elements are invalidated by · 1ba34bf8

kmillikin@chromium.org authored Mar 24, 2009

InvalidateFrameSlotAt.

Review URL: http://codereview.chromium.org/53008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1594 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

1ba34bf8

Add pointers to a VirtualFrame that keep track of the indexes of register · bbd21a59

whesse@chromium.org authored Mar 24, 2009

frame elements.
Review URL: http://codereview.chromium.org/53005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1593 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

bbd21a59

Preparing push of version 1.1.2 to trunk. · efba01b8

ager@chromium.org authored Mar 24, 2009

Review URL: http://codereview.chromium.org/42556

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1592 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

efba01b8

Add newline at end of test-debug.cc file. TBR=sgjesse@chromium.org · dd95e1d2
kasperl@chromium.org authored Mar 24, 2009
```
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1591 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
```
dd95e1d2

Added a test for debugger agent protocol message. · 1bf46b6b

sgjesse@chromium.org authored Mar 24, 2009

Review URL: http://codereview.chromium.org/42555

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1590 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

1bf46b6b

Test a few assertions that should hold. · c8c5d982

lrn@chromium.org authored Mar 24, 2009


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

c8c5d982

Change the debugger agent protocol to be RFC-822 compilant. · cf8c7cc3

sgjesse@chromium.org authored Mar 24, 2009

All headers are now terminated with CRLF and not jusr LF. The beginning of the body is indicated by an empty header (just CRLF).
Review URL: http://codereview.chromium.org/42553

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1588 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

cf8c7cc3

Fix ARM performance regression on constructing array · 70aa1fde

kasperl@chromium.org authored Mar 24, 2009

and object literals by using the same logic as we use
on IA-32.
Review URL: http://codereview.chromium.org/53001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1587 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

70aa1fde

Update test expectation. · 4e747256

kmillikin@chromium.org authored Mar 24, 2009

TBR=plesner

Review URL: http://codereview.chromium.org/42554

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1586 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

4e747256

Fix issue 284. · 5e2f3adf

kmillikin@chromium.org authored Mar 24, 2009

The problem was continuing out of the body of a for/in (where we do
register allocation) to the loop update (where we do not). Variables
allocated to registers where not preserved.

Review URL: http://codereview.chromium.org/53002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1585 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

5e2f3adf

Fix bug in context stores, where the store was ignored if there was a · d4e66137

ager@chromium.org authored Mar 24, 2009

read-only property in the prototype chain. Stores should only be
ignored if there is a read-only property on the object itself.
Review URL: http://codereview.chromium.org/42508

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1584 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

d4e66137

Add a new "pseudo" platform for POSIX. It it will contain code shared by the... · 96088bab

sgjesse@chromium.org authored Mar 24, 2009

Add a new "pseudo" platform for POSIX. It it will contain code shared by the platforms supporting POSIX. Currently Linux, Mac OS and FreeBSD are considered POSIX platforms using the POSIX implementation.

Initially the platform socket implementation has been moved from the Linux, Mac OS and FreeBSD platforms to the POSIX platform.

If implementation which ends up in the POSIX platform at some point cannot be the same across POSIX platforms the implementation can easily be moved back into the individual POSIX "complient" platforms.
Review URL: http://codereview.chromium.org/51001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1583 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

96088bab

Marked bug 284 as passing on arm. · dbe4e4a6

christian.plesner.hansen authored Mar 24, 2009

Review URL: http://codereview.chromium.org/49013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1582 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

dbe4e4a6

23 Mar, 2009 10 commits

Added test case for issue 284. · cff1d271

christian.plesner.hansen authored Mar 23, 2009

Review URL: http://codereview.chromium.org/52031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1581 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

cff1d271

Fixed Linux compiler warning. · b5d265a3

sgjesse@chromium.org authored Mar 23, 2009

TBR=erik.corry@gmail.com
Review URL: http://codereview.chromium.org/49010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1580 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

b5d265a3

Extend debugger agent protocol with a connect message.Added a name of the... · d8e53cf1

sgjesse@chromium.org authored Mar 23, 2009

Extend debugger agent protocol with a connect message.Added a name of the embedding application when enabeling the debugger agent.Send a connection message from the debugger agent to the remote debugger when connecting. This message contains the V8 version, the protcol version (currently 1) and the name of the embedding application. Currently this information is just printed raw as received.
Review URL: http://codereview.chromium.org/52012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1579 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

d8e53cf1

Add a new C++ pointer wrapping API to External to not dilute the · 7b50c072

iposva@chromium.org authored Mar 23, 2009

External::Cast.

Review URL: http://codereview.chromium.org/52021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1578 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

7b50c072

Fix flaw in VirtualFrame::SetElementAt handling multiple copies of elements. · 34f71b11

whesse@chromium.org authored Mar 23, 2009

Review URL: http://codereview.chromium.org/47006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1577 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

34f71b11

Fix build · 010443b9

whesse@chromium.org authored Mar 23, 2009

TBR=kmillikin@chromium.org
Review URL: http://codereview.chromium.org/52013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1576 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

010443b9

Change VirtualFrame::AdjustCopies to mark target as invalid. · ba16099e

whesse@chromium.org authored Mar 23, 2009

Change its name to VirtualFrame::InvalidateFrameSlot
Review URL: http://codereview.chromium.org/50012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1575 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ba16099e

Reapply portability fixes, this time without 64 bit system bug. · aa81281b

erik.corry@gmail.com authored Mar 23, 2009

Review URL: http://codereview.chromium.org/52010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1574 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

aa81281b

Profiler ticks every 1ms instead of every 10ms. · 5fb08f90

erik.corry@gmail.com authored Mar 23, 2009

Review URL: http://codereview.chromium.org/42498

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1573 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

5fb08f90

Fix bug in regexp the breaks interpreter. · 9ca7c8af

lrn@chromium.org authored Mar 23, 2009


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1572 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

9ca7c8af