To avoid deopts a few extra changes were needed:
o Enable megamorphic state for special property loads on
primitives. We used to flip between monomorphic stubs.
o Extract pure string (no string wrapper support) version of the
string length stub.
Review URL: http://codereview.chromium.org/6334015
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6472 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
conversions for external array types, which implement the Typed Array
spec. (Revision of http://codereview.chromium.org/6315004 .)
Prefer SSE2 code path on x86 processors. Non-SSE2 processors now make
a slow runtime call for float-to-int conversions. Use SSE3 for 32-bit
signed and unsigned int array types where possible.
The movement of code from ic-arm.cc to stub-cache-arm.cc caused the
VFP3 code path to be tested for the first time. Fixed bugs in the
register usage and in the constant value stored into integer arrays
for NaN and +/-Infinity.
Added new truncation test to test-api.cc. Storage of NaN and +/-Inf
was already covered. Ran unit tests on x86, x64 and ARM simulator.
Tested ia32 and x64 code in Chromium on Mac and Linux respectively
with Typed Array unit tests and WebGL content.
BUG=http://code.google.com/p/chromium/issues/detail?id=50972
TEST=test-api/ExternalArrays
Review URL: http://codereview.chromium.org/6303012
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6431 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Unfortunately, arguments is pretty much the normal JS object. For now
I am adding more sanity checks (in hope that typically arguments
list is rather short.) However it probably requires more systematic
treatment, for example, we could optimistically copy elements until
we meet first hole and in this case resort to JS builtin.
Review URL: http://codereview.chromium.org/6062006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6256 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This removes the expected failure of a mozilla test that we now pass
(an empty array is returned if Array.splice is called with no
arguments) and fixes debug check failure by allocating a new empty
array using AllocateEmptyArray if splice is called with zero arguments
(makes sure we can always create a handle).
Review URL: http://codereview.chromium.org/3218010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5377 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Object model changes
----------------------------------------
New fixed_cow_array_map is used for the elements array of a JSObject
to mark it as COW. The JSObject's map and other fields are not
affected. The JSObject's map still has the "fast elements" bit set. It
means we can do only the receiver map check in keyed loads and the
receiver and the elements map checks in keyed stores. So introducing
COW arrays doesn't hurt performance of these operations. But note that
the elements map check is necessary in all mutating operations because
the "has fast elements" bit now means "has fast elements for reading".
EnsureWritableFastElements can be used in runtime functions to perform
the necessary lazy copying.
Generated code changes
----------------------------------------
Generic keyed load is updated to only do the receiver map check (this
could have been done earlier). FastCloneShallowArrayStub now has two
modes: clone elements and use COW elements. AssertFastElements macro
is added to check the elements when necessary. The custom call IC
generators for Array.prototype.{push,pop} are updated to avoid going
to the slow case (and patching the IC) when calling the builtin should
work.
COW enablement
----------------------------------------
Currently we only put shallow and simple literal arrays in the COW
mode. This is done by the parser.
Review URL: http://codereview.chromium.org/3144002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The IC stub is completely generic, so there will only be one such stub
in the system.
Added a new overloaded version of the macro assembler RecordWrite
method for cases where we have the address we store to computed up
front.
Review URL: http://codereview.chromium.org/2804029
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4991 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
A potential issue with this change is creating lots of maps when
objects flip between fast/slow elements modes. We could add special
transitions to avoid this. Yet testing this on our benchmarks, gmail,
and wave seems to indicate that this is not a real problem.
Review URL: http://codereview.chromium.org/2870018
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4941 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Added support for more precise break points when debugging and stepping. To achieve that additional nop instructions are inserted where breaking would otherwise be impossible. The number of nop instructions inserted are sufficient to make place for patching with a call to a debug break code stub. On Intel that is 5 nop's for 32-bit and 13 for 64-bit. Om ARM 3 nop instructions (12 bytes) are required.
In order to avoid inserting nop's in to many places a simple ast checker have been added to check whether there are breakable code in a statement or expression. If it is possible to break in an expression no additional break enabeling code is inserted.
Added break locations to the true and false part of a conditional expression.
Added stepping tests to cover more constructs.
These changes are only in the full compiler.
Changed the default value for the option --debugger in teh d8 shell from true to false. The reason for this is that with --debugger turned on the full compiler will be used for all code in when running d8, which can be unexpeceted.
Review URL: http://codereview.chromium.org/2693002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4820 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
- New сardmarking write barrier handles large objects and normal objects in a similar fashion (no more additional space for pointer tracking is required, no conditional branches in WB code).
- Changes to enable oldspaces iteration without maps decoding:
-- layout change for FixedArrays: length is stored as a smis (initial patch by
Kevin Millikin)
-- layout change for SharedFunctionInfo: integer fields are stored as smi on
arm, ia32 and rearranged on x64.
-- layout change for String: meaning of LSB bit is fliped (1 now means hash not
computed); on x64 padding is added.
-- layout of maps is _not_ changed. Map space is currently iterated in a special
way.
Review URL: http://codereview.chromium.org/2144006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4715 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
-- layout change for FixedArrays: length is stored as a smis (initial patch by Kevin Millikin)
-- layout change for SharedFunctionInfo: integer fields are stored as smi on arm, ia32 and rearranged on x64.
-- layout change for String: meaning of LSB bit is fliped (1 now means hash not computed); on x64 padding is added.
-- layout of maps is _not_ changed. Map space is currently iterated in a special way.
- Cardmarking write barrier. New barrier handles large objects and normal objects in a similar fashion (no more additional space for pointer tracking is required, no conditional branches in WB code).
Review URL: http://codereview.chromium.org/2101002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4685 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
function instead of baking in the address of the first one that we see
in code.
This removes the need for fixups processing and makes the stubs safe
when there is no natives cache and therefore multiple versions of the
builtin functions.
Review URL: http://codereview.chromium.org/594009
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3832 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
1. Avoid using SKIP_WRITE_BARRIER when we don't have to (smis).
2. Check and document the remaining uses of SKIP_WRITE_BARRIER.
3. Only allow GetWriteBarrierMode when in an AssertNoAllocation scope.
The only functional change should be in DeepCopyBoilerplate where we
no longer use the write barrier mode (because of allocations). I'm
running benchmarks to see if this has a measurable impact on performance.
Review URL: http://codereview.chromium.org/558041
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3743 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
in particular).
* Called function is passed on the stack instead of
using a static variable.
* Builtins that don't need the called function don't
get it.
* Made is_construct statically known to HandleApiCall
by setting custom construct stub for API functions.
Review URL: http://codereview.chromium.org/536065
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3613 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Instead of going through a runtime function for keyed loads
on strings we invoke a separate specialized stub that
assumes string as receiver type and the key to be a number.
The stub calls a JS builtin function to return the corresponding one-character string.
Review URL: http://codereview.chromium.org/521041
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3556 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
specification under development. The optimizations are patterned after
those previously done for CanvasPixelArray. This CL adds all of the
necessary framework but continues to use the generic KeyedLoadIC and
KeyedStoreIC code, to create a baseline for benchmarking purposes. The
next CL will add the optimized ICs to ic-ia32.cc and ic-x64.cc.
These new CanvasArray types have different semantics than
CanvasPixelArray; out-of-range values are clamped via C cast
semantics, which is cheaper than the clamping behavior specified by
CanvasPixelArray. Out-of-range indices raise exceptions instead of
being silently ignored.
As part of this work, pulled FloatingPointHelper::AllocateHeapNumber
up to MacroAssembler on ia32 and x64 platforms. Slightly refactored
KeyedLoadIC and KeyedStoreIC. Fixed encoding for fistp_d on x64 and
added a few more instructions that are needed for the new ICs. The
test cases in test-api.cc have been verified by hand to exercise all
of the generated code paths in the forthcoming specialized ICs.
Review URL: http://codereview.chromium.org/293023
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3096 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The construction of arrays when using the the Array function either as a constructor or a normal function is now handled fully in generated code in most cases. Only when Array is called with one argument which is either negative or abowe JSObject::kInitialMaxFastElementArray (which is currently 1000) or if the allocated object cannot fit in the room left in new space is the runtime system entered.
Two new native code built-in functions are added one for normal invocation and one for the construct call. The existing C++ builtin is renamed, but kept. When the normal invocation cannot be handled in generated code the C++ builtin is called. When the construct invocation cannot be handled in native code the generic construct stub is called (which will end up in the C++ builtin through a construct trampoline).
One thing that might be changed is preserving esi (constructor function) during the handling of a construct call. We know precisily what function we where calling anyway and can just reload it. This could remove the parameter construct_call to ArrayNativeCode and remove the handling of this from that function.
The X64 and ARM implementations are not part of this changelist.
Review URL: http://codereview.chromium.org/193125
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2899 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Got rid of the debug break on return entry code which did not add anything. It just jumped directly to the debug break on return code.
Removed the CodePatcher class on x64 as it was not implemented.
Added instruction cache flush to where the return sequence was patched on x64.
Added some missing ENABLE_DEBUGGER_SUPPORT #ifdef/#endif.
Review URL: http://codereview.chromium.org/193057
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2863 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Changed CalledAsConstructor to process the raw runtime stack directly,
rather than using a StackFrameIterator. Using an interator turns out
to be quite expensive and the vast majority of the work done is not
relevant to deciding if we've been called as a constructor.
This speeds up getElementById by ~13%.
Review URL: http://codereview.chromium.org/160325
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2578 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This is a trivial per-row compression:
- short aliases are introduced for events and code creation tags;
- in tick events, offsets are used instead of absolute addresses;
- removed 'code-allocation' event, as it seems not used.
The first two options are depend on the new flag: 'compress-log', which is off by default.
On benchmarks run w/o snapshot, this gives 45% log size reduction.
Review URL: http://codereview.chromium.org/119304
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2122 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
called from within a loop or not. In the past we lost the
information if a call site went megamorphic before a lazily
compiled callee was called for the first time. Now we track
that correctly (this is an issue that affects richards).
We still don't manage to track the in-loop state through a
constructor call, since constructor calls use LoadICs instead
of CallICs. This issue affects delta-blue. So in this patch
we assume that lazy compilations that don't happen through a
CallIC happen from inside a loop. I have an idea to fix this
but this patch is big enough already.
With our improved tracking of in-loop state I have switched
off the inlining of in-object loads for code that is not in
a loop. This benefits compile speed. One issue is that
eagerly compiled code now doesn't get the in-object loads
inlined. We need to eagerly compile less code to fix this.
Review URL: http://codereview.chromium.org/115744
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2046 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
ENABLE_DEBUGGER_SUPPORT is enabled by default unless it is on Android platform.
On Android platform, it can also enabled by passing -DENABLE_DEBUGGER_SUPPORT flag to the compiler.
This should not affect any existing build (I hope, cross my fingers) except the build in real Android environment (in other word, it only affects me now).
There are lot of room for code refactoring in stead of using #ifdef all over the place. I will leave this to v8 folks.
Review URL: http://codereview.chromium.org/77035
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1745 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
operation. I have another version of this change that does not remove the
special inline caches for difference sized strings. The other version is ever
so slightly faster, but the nice thing about this version is that it removes
253 lines of code.
Review URL: http://codereview.chromium.org/8187
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@603 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Here is a description of the background and design of split window in Chrome and V8:
https://docs.google.com/a/google.com/Doc?id=chhjkpg_47fwddxbfr
This change list splits the window object into two parts: 1) an inner window object used as the global object of contexts; 2) an outer window object exposed to JavaScript and accessible by the name 'window'. Firefox did it awhile ago, here are some discussions: https://wiki.mozilla.org/Gecko:SplitWindow. One additional benefit of splitting window in Chrome is that accessing global variables don't need security checks anymore, it can improve applications that use many global variables.
V8 support of split window:
There are a small number of changes on V8 api to support split window:
Security context is removed from V8, so does related API functions;
A global object can be detached from its context and reused by a new context;
Access checks on an object template can be turned on/off by default;
An object can turn on its access checks later;
V8 has a new object type, ApiGlobalObject, which is the outer window object type. The existing JSGlobalObject becomes the inner window object type. Security checks are moved from JSGlobalObject to ApiGlobalObject. ApiGlobalObject is the one exposed to JavaScript, it is accessible through Context::Global(). ApiGlobalObject's prototype is set to JSGlobalObject so that property lookups are forwarded to JSGlobalObject. ApiGlobalObject forwards all other property access requests to JSGlobalObject, such as SetProperty, DeleteProperty, etc.
Security token is moved to a global context, and ApiGlobalObject has a reference to its global context. JSGlobalObject has a reference to its global context as well. When accessing properties on a global object in JavaScript, the domain security check is performed by comparing the security token of the lexical context (Top::global_context()) to the token of global object's context. The check is only needed when the receiver is a window object, such as 'window.document'. Accessing global variables, such as 'var foo = 3; foo' does not need checks because the receiver is the inner window object.
When an outer window is detached from its global context (when a frame navigates away from a page), it is completely detached from the inner window. A new context is created for the new page, and the outer global object is reused. At this point, the access check on the DOMWindow wrapper of the old context is turned on. The code in old context is still able to access DOMWindow properties, but it has to go through domain security checks.
It is debatable on how to implement the outer window object. Currently each property access function has to check if the receiver is ApiGlobalObject type. This approach might be error-prone that one may forget to check the receiver when adding new functions. It is unlikely a performance issue because accessing global variables are more common than 'window.foo' style coding.
I am still working on the ARM port, and I'd like to hear comments and suggestions on the best way to support it in V8.
Review URL: http://codereview.chromium.org/7366
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@540 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
ExtendStorage did not work with keyed store IC.
- Reduced instructions generated when performing a tail call to
kSharedStoreIC_ExtendStorage
- Moved test/mjsunit/bugs/bug-109.js
to test/mjsunit/keyed-storage-extend.js
Review URL: http://codereview.chromium.org/6526
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@455 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
- Simplified frame entry and frame exit code.
- Added ArgumentsAdaptorTrampoline and check for matching argument counts in the InvokePrologue.
- Removed definition and uses of USE_OLD_CALLING_CONVENTIONS.
- Changed MacroAssembler::InvokeBuiltin to match ia32 version.
- Start introducing convenience instructions in the ARM assembler as needed. These instructions take all Register parameters to avoid extra typing of "Operand(reg)".
To keep the architectures in sync these changes have been made to the ia32 files:
- Changed MacroAssembler::EnterFrame(StackFrame::Type type) to MacroAssembler::EnterInternalFrame().
These parts are still missing:
- unimplemented: Builtins::Generate_FunctionApply - large limit
- unimplemented: Builtins::Generate_ArgumentsAdaptorTrampoline - non-function call
- The files have not been lint'd yet.
Review URL: http://codereview.chromium.org/1930
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@289 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Added better test support.
Added load, quit and version functions to the shell sample so it's easier to run benchmarks and tests.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Cleaned up ARM version by removing top of stack caching and by introducing push/pop elimination.
Cleaned up the way runtime functions are called to allow runtime calls with no arguments.
Changed Windows build options to make sure that exceptions are disabled and that optimization flags are enabled.
Added first version of Visual Studio project files.
git-svn-id: http://v8.googlecode.com/svn/trunk@13 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
