As the start index is already passed it is easy to calculate the "at start" boolean in generated code. Also as direct entry has been implemented this needs to be done in generated code anyway, and therefore might as well be moved to the generated code for RegExp. The "at start" value is now calcualted as a local variable on the native RegExp frame based on the value of the start index argument.
The x64 version have been tested on both Linux and 64-bit Windows Vista.
For ARM I have tested cctest/test-regexp on ARM hardware, but the rest of the tests have only been run on the ARM simulator.
Review URL: http://codereview.chromium.org/554078
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3709 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
generator.
Contexts are no longer stored in the AST but in the code generator's
state. This means that the running the code generator selector is not
required to use the toplevel code generator (for instance, if we
already know that we can and should use it).
Review URL: http://codereview.chromium.org/546075
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3645 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
left and right sides of the comparison to reflect reality. Don't
check explicitly for NaNs when using VFP3 since the compare
operation can signal this case with the v flag. Use cmp instead
of tst in the fast compilers since tst leaves the v flag unchanged
and thus can only work by accident on non-equality comparisons.
Review URL: http://codereview.chromium.org/551048
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3625 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When a function is called with a value type as the receiver this is now boxed as an object.
This is a low-impact solution where the receiver is only boxed when required. For IC calls to the V8 builtins values are not boxed and as most of the functions on String.prototype, Number.prototype and Boolean.prototype are sitting there most IC calls on values will not need any boxing of the receiver.
For calls which are not IC calls but calls through the CallFunctionStub a flag is used to determine whether the receiver might be a value and only when that is the case will the receiver be boxed.
No changtes to Function.call and Function.apply - they already boxed values. According to the ES5 spec the receiver should not be boxed for these functions, but current browsers have not adopted that change yet.
BUG=223
TEST=test/mjsunit/value-wrapper.js
TEST=test/mjsunit/regress/regress-crbug-3184.js
Review URL: http://codereview.chromium.org/542087
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3617 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
* Refactor VisitProperty to use the platform-specific methods for emitting the IC calls.
* Refactor recording of source positions in the top-level compiler.
* Correct the recorded source positions for assignments and property loads.
* Fix bug on x64 where source positions were not recorded before a calling a call-IC.
* Correct some inconsistencies between IA-32 and X64 top-level code generator.
We now pass all regression tests with
--always-fast-compiler.
Review URL: http://codereview.chromium.org/550043
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3612 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Replace runtime call to NumberAdd with call to binary op stub.
Until now the top-level compiler always called a runtime function
for count operations.
In some places we expected in the JS builtins smis as arguments.
If we perform a count operation before all smis would get converted into
heap numbers by the runtime number add function and result in a runtime
assert.
Also: Add missing debugger information in the top-level compiler for
do-while loops.
Review URL: http://codereview.chromium.org/548029
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3610 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When initializing the special local variable containing the reference to the enclosing
function in named functions we now (correctly) emit an INIT_CONST instead of INIT_VAR,
and we correctly bail out in the top-level code generator.
Also part of this change is adding missing statement position information
for some statements in the top-level code generator.
Review URL: http://codereview.chromium.org/536029
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3587 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
for partial snapshots. After reserving space we can be sure that allocations
will happen linearly (no GCs and no free-list allocation). This change also
contains the start of the partial snapshot support, which, however is not yet
completed or tested.
Review URL: http://codereview.chromium.org/545026
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3584 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
instruction sequence for jumps:
mov(ip, Operand(target, rmode), LeaveCC, cond);
bx(ip, cond)
Changed a JS call in the compare stub to a tail call to avoid GC
problems where the pushed return address is not updated on GC.
Review URL: http://codereview.chromium.org/549022
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3582 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
* Rename the operation of filling a context with a value. Formerly it
was 'Move', now it's 'Apply' so as to avoid confusion with various
other Moves (eg, in the toplevel codegen, in the macro assemblers).
* Use the abstraction Drop rather than math on the stack pointer.
* Add a predicate on AST expression nodes to abstract a repeated test
whether we should us a regular (named) IC or a keyed IC.
Review URL: http://codereview.chromium.org/546006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3580 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Added a stub to allocate and fill a string object with a substring from another string.
Use the rep movs instruction to copy the string data as it turned out to be the fastest way.
While preparing this I experimented with some SSE2 instructions, so the instructions movdqa and movdqu are still in the IA-32 assembler even though they are not used.
Review URL: http://codereview.chromium.org/525085
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3554 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Calls to RegExp no longer have to be via a call to the runtime system. A new stub have been added which can handle this call in generated code. The stub checks all the parameters and creates RegExp entry frame in the same way as it is created by the runtime system. Bailout to the runtime system is done whenever an uncommon situation is encountered or when the static data used is not initialized. After running the native RegExp code the last match info is updated like in the runtime system.
Currently only ASCII strings are handled.
Added another argument to the RegExp entry frame. It indicated whether the call is direct from JavaScript code or through the runtime system. This information is used when RegExp execution is interrupted. If an interruption happens when RegExp code is called directly a retry is issued causing the interruption to be handled via the runtime system. The reason for this is that the direct call to RegExp code does not support garbage collection.
Review URL: http://codereview.chromium.org/521028
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3542 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
generator.
The Slot::Type enumeration has four values. It should never be
necessary to use a default to handle the case of a value out of range
of the enumeration. Doing so silences a useful warning when one of
the enumeration values is actually forgotten or when a new enumeration
value is added.
Review URL: http://codereview.chromium.org/521019
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3532 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Until now we only supported postfix operations on global variables.
This change add generic count operations to the top-level compiler.
I tried to re-use code from the code generator used for assignment expressions
where possible.
Review URL: http://codereview.chromium.org/496009
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3530 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
If a function contains more than a certain number of locals (IA32: 9, X64: 6, ARM: 4)
a loop for initializing the locals with 'undefined' is more compact.
For less locals we unroll that loop by emitting a sequence of push instructions.
Review URL: http://codereview.chromium.org/515012
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3521 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When generating code for object and array literals we performed
the check if the a boilerplate already exists in generated code.
In the top-level compiler we now do this check in a new runtime
function. This makes the generated code more compact for top-level code.
Review URL: http://codereview.chromium.org/465148
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3437 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
1. Refactor the structure of VisitAssignment: The existing code is not ideal
to be extended with support for compound assignments.
2. Reuse common code for keyed property assigments: Now variables rewritten
to a property (.arguments access) are treated like normal keyed property
assignments. This allows us to remove some code duplication.
Review URL: http://codereview.chromium.org/456024
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3425 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This adds a code stub which can do most of what Heap::AllocateConsString can do. It bails out if the result cannot fit in new space or if the result is a short (flat) string and one argument is an ascii string and the other a two byte string. It also bails out if adding two one character strings as Heap::AllocateConsString has special handling of this utilizing the symbol table. The stub is used both for the binary add operation and for StringAdd calls from runtime JavaScript files. Extended the string add test to cover all sizes of flat result stings.
Review URL: http://codereview.chromium.org/442024
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3400 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The different length string types was used to encode the string length and the hash in one field. This is now split into two fields one for length and one for hash. The hash field still encodes the array index of the string if it has one. If an array index is encoded in the hash field the string length is added to the top bits of the hash field to avoid a hash value of zero.
On 32-bit this causes an additional 4 bytes to be used for all string objects. On 64-bit this will be half on average dur to pointer alignment.
Review URL: http://codereview.chromium.org/436001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3350 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Summary:
This change fixes a performance regression introduced by the special
handling of regular expressions in typeof expressions.
As a result we regain ~8% speedup on 3d-raytrace and ~13% on boyer
(vs bleeding edge)
Description:
The macros IS_OBJECT and IS_FUNCTION are frequently used in the
JS runtime functions.
By introducing new inlined runtime functions %_IsFunction and %_IsObject
we avoid invoking the more expensive %_ClassOf function plus comparing
its result to a string.
Review URL: http://codereview.chromium.org/399111
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3335 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Condition block of do/while statements is a valid break location so it should have its own position. The block is represented by a regular Expression node so we cannot store the position in it, instead the position is stored in a separate field in DoWhileStatement AST node.
BUG=514
Review URL: http://codereview.chromium.org/385136
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3312 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
available. We use this to ensure that snapshots on MacOSX can
use SSE2 instructions. Also clean up and assertify the
handling of things we can't do when we are generating a
snapshot. Fix a bug in the new serialization tests where
they activated Snapshot::enable() too late after code had been
generated that assumed no snapshots.
Review URL: http://codereview.chromium.org/391051
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3301 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
regular expressions to be consistent with typeof in other contexts.
Typeof regular expressions should now be 'function' in all contexts.
In the JS natives, IS_FUNCTION, IS_OBJECT, and IS_REGEXP return the
same answers as before.
Review URL: http://codereview.chromium.org/360053
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3231 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
In the generated code for function.apply there was a loop checking the stack limit for interruption. This loop would call into the runtime system to handle interuption and keep running until there was no interruption. However if the interuption was debug break the runtime system would never clear the interruption as debug break is prevented in builtins are prevented and the assumption here was that returning with the debug break flag set would move execution forward.
Renamed initial_jslimit and initial_climit to real_jslimit and real_climit. Renamed a few external references related to the stack limit as well.
Exposed the real stack limit to generated code to make the stack check when entering function.apply use the real stack limit and not the stack limit which is changed to signal interruption.
Added the real stack limit to the roots array.
BUG=http://code.google.com/p/v8/issues/detail?id=493
TEST=cctest/test-debug/DebugBreakFunctionApply
Review URL: http://codereview.chromium.org/345048
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3229 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
separate JS stack.
In exception handling, we need to be able to compare addresses into
the JavaScript portion of the stack with the address of a C++ handler
on the stack. Since the stacks are separate on the simulator, we need
a JavaScript stack address corresponding to a C++ try catch handler in
order to perform valid address comparisons.
On the simulator, we now link the C++ try catch handlers indirectly
through the JS stack and use the JS stack indirection address for
comparisons.
JS C++
handler
[C++ address] <------ next_
\
\
\----> handler
[C++ address] <------ next_
On actual hardware the C++ try catch handlers continue to be directly
linked.
BUG=http://code.google.com/p/v8/issues/detail?id=271
Review URL: http://codereview.chromium.org/360004
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3228 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
There were two separate implementations of the function
'BuildBoilerplate' that is used to compile function declarations and
function literals. The implementations did not do exactly the same
thing. In particular, one ignored the flag --lazy.
Combine the two implementations.
Review URL: http://codereview.chromium.org/360011
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3218 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
used to signal that an expression was the immediate subexpression of
typeof, or (?) in the arm of a conditional expression itself in the
typeof state.
It was inconsistently consulted. It was not used for property loads,
but only for slot loads. This means that we matched the Webkit JSC
(not Spidermonkey) behavior for:
typeof(true ? x : y) // throws ReferenceError
and we matched the SpiderMonkey behavior (not JSC) for:
with ({}) { typeof(true ? x : y) } // ==> "undefined"
Now we are expected to match the JSC behavior in all cases.
Review URL: http://codereview.chromium.org/362004
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3212 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
directly-applied function literals that are themselves compiled with
the top-level code generator.
The choice is guarded by a test that the function is anonymous (thus
not expected to be recursive) and not in a loop.
A compilation hint is set in the shared function info and used to make
the choice.
Review URL: http://codereview.chromium.org/341081
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3206 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
context. Test contexts are used for the left subexpressions of
short-circuited boolean operators. The right subexpressions inherit
their expression context from the binary op expression.
Compilation of short-circuited operations in effect and test context
is straightforward:
effect(e0 || e1) =
test(e0, L0, L1)
L1: effect(e1)
L0:
test(e0 || e1, L0, L1) =
test(e0, L0, L2)
L2: test(e1, L0, L1)
Because the value of the first subexpression may be needed as the
value of the whole expression in a value context, we introduce a
hybrid value/test contest (the value is needed if true, but not if
false).
value(e0 || e1) =
value/test(e0, L0, L1)
L1: value(e1)
L0:
The compilation of value/test and test/value (introduced by boolean
AND) is:
value/test(e0 || e1, L0, L1) =
value/test(e0, L0, L2)
L2: value/test(e1, L0, L1)
test/value(e0 || e1, L0, L1) =
test(e0, L0, L2)
L2: test/value(e1, L0, L1)
Boolean AND is the dual. The AST nodes themselves (not their parents)
are responsible for producing the proper result (effect, value, or
control flow) depending on their context.
Review URL: http://codereview.chromium.org/339082
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3187 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
o.x() and o[expr]()
other changes:
- Fix missing relocation info for StoreIC on global object.
- Generate only one common return sequence instead of always appending
"return <undefined>" at the end of each function: The first JS
return-statement will generate the common return sequence. All
other return-statements will generate a unconditional branch to the common
return sequence.
Review URL: http://codereview.chromium.org/340037
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3183 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The code for .result = (b.y = 99) where b is a global variable is:
push [esi+0x17]
mov ecx,0xf5c229ad ;; object: 0xf5c229ad <String[1]: b>
call LoadIC_Initialize
nop
mov [esp],eax
mov eax,0xc6
mov ecx,0xf5c25c41 ;; object: 0xf5c25c41 <String[1]: y>
call StoreIC_Initialize
nop
mov [esp],eax
pop [ebp+0xf4]
There is still some room for improvement in the generated code.
Other changes:
- Replaced switch-statement in FastCodeGenerator::VisitProperty with DropAndMove(...)
- Do not emit nop after IC calls on ARM.
Review URL: http://codereview.chromium.org/347001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3180 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
context of the expressions they label. Introduce an "unintialized"
location to catch failure to assign any location at all.
Changed the object literal initialization on ARM to use a Store IC in
the same cases where it did on the other platforms. This was required
because the location of the literal property name is given an
"unitialized" location.
Review URL: http://codereview.chromium.org/339045
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3171 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
location to an Expression AST node from the node's parent to the node
itself.
This allows an inherited code generation context from a parent node to
be passed arbitrarily far down the tree (eg, the subexpression of a
unary not is in the same context as the unary expression itself, the
then and else subexpressions of the ternary operator are in the same
context as the whole expression, and so forth).
We do not yet take advantage of this in the backend (eg, the right
subexpression of short-circuited OR is still compiled by using the
parent's destination location, rather than the subexpression's
itself).
Review URL: http://codereview.chromium.org/340005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3163 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Literals now have a location of temporary by default and are
responsible for moving themselves into their location like all other
expressions.
The constant location turned out not to allow us to avoid checking
subexpressions in AST interior nodes, and it turned out to require
checking after some normal calls to Visit (like for the arguments to a
call). With this change do not have to check after a call to Visit
that we got our result in the expected location.
Review URL: http://codereview.chromium.org/339004
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3137 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
of individual changes:
- Added infrastructure for custom stub caching.
- Push the code object onto the stack in exit calls instead of a
debug/non-debug marker.
- Remove the DEBUG_EXIT frame type.
- Add a new exit stub generator for API getters.
Review URL: http://codereview.chromium.org/330017
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3130 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
I also added more unit tests for literals.
Right now, the fast compiler produces code very similar to
the existing code generator. We may consider different ways to
further compact the generated code for top-level code.
ARM always goes through a runtime function to initialize computed
properties in an object literal whereas IA32 and x64 use StoreIC.
Review URL: http://codereview.chromium.org/316009
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3129 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
As the list of code-stubs is used in two places it is now handled through a macro to keep this in sync. As some code-stubs is only used on ARM the list have been split into two parts to indicate this and get rid of dummy implementation on ia32 and x64 platforms.
BUG=484
Review URL: http://codereview.chromium.org/335025
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3127 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
in a non-test (ie, value or effect) context. (It is implicitly not in
a test context because the code generator does not support expressions
in a test context yet.)
Compilation is essentially the same as in the optimized code
generator. The expression (e0 || e1) is compiled as if it were
(let (temp = e0) temp ? temp : e1).
On ia32 and x64 a single shared ToBoolean stub is used to convert a
value to a flag. The inlined checks assumed by the stub are reordered
to compare to undefined (the common case in toplevel code?) first. On
ARM a call to the runtime is used. In the interest of code size no
checks are yet inlined on ARM.
Review URL: http://codereview.chromium.org/334006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3118 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
currently compiled the same as with the optimizing compiler: they are
cloned from a boilerplate object and the boilerplate objects are
lazily constructed.
Also changed argument pushing on ARM to use stm (store multiple),
which required changing the order of arguments to the runtime
functions DeclareGlobals and NewClosure. They were only used from
generated code.
Finally, changed the toplevel code generator so that stack pops to
discard a temporary became addition to the stack pointer on ia32 and
x64.
Review URL: http://codereview.chromium.org/303021
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3110 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
For .result = /abc.*/g we generate the following IA-32 code:
...
mov ebx,[edi+0x17]
mov eax,[ebx+0xb]
cmp eax, 0xf5d0e135 ;; object: 0xf5d0e135 <undefined>
jnz done
push ebx
push 0x2
push 0xf5d13805 ;; object: 0xf5d13805 <String[5]: abc.*>
push 0xf5d13815 ;; object: 0xf5d13815 <String[1]: g>
call RuntimeStub_MaterializeRegExpLiteral
done:
push eax
pop [ebp+0xf4]
...
This is very similar to the code previously generated except we do not
generate deferred code for the case where we call the runtime.
On ARM we use the stm instruction to make pushing the arguments more compact.
Review URL: http://codereview.chromium.org/300037
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3109 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
constant known at compile time. Do not ever use the stack to
materialize (non-function-argument) constants. Currently, constants
are only the non-materialized, non-function literals in the AST.
It is a known issue that there is no test coverage for the cases of
assigning a non-literal to a variable and returning a literal. Those
code paths are unreachable and tests will be added when they become
reachable.
For the code '.result = true', we had previously on ia32:
27 push 0xf5c28161 ;; object: 0xf5c28161 <true>
32 pop [ebp+0xf4]
Now:
27 mov eax,0xf5c26161 ;; object: 0xf5c26161 <true>
32 mov [ebp+0xf4],eax
======== We had previously on x64:
25 movq r10,0x7fb8c2f78199 ;; object: 0x7fb8c2f78199 <true>
35 push r10
37 pop [rbp-0x18]
Now:
25 movq r10,0x7fb131386199 ;; object: 0x7fb131386199 <true>
35 movq [rbp-0x18],r10
The generated code for ARM did not include the extra memory traffic.
It was already eliminated by the ARM assembler's push/pop elimination.
Review URL: http://codereview.chromium.org/300003
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3088 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Slots appear only indirectly in the AST (through variables linked to
variable proxies). Slots are shared among variable references, so
putting compilation-time state on them is potentially a source of
bugs. Avoid it for now.
Review URL: http://codereview.chromium.org/284009
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3079 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
fast-mode code generator.
AST expression nodes are annotated with a location when doing the
initial syntactic check of the AST. In the current implementation,
expression locations are 'temporary' (ie, allocated to the stack) or
'nowhere' (ie, the expression's value is not needed though it must be
evaluated for side effects).
For the assignment '.result = true' on IA32, we had before (with the
true value already on top of the stack):
32 mov eax,[esp]
35 mov [ebp+0xf4],eax
38 pop eax
Now:
32 pop [ebp+0xf4]
======== On x64, before:
37 movq rax,[rsp]
41 movq [rbp-0x18],rax
45 pop rax
Now:
37 pop [rbp-0x18]
======== On ARM, before (with the true value in register ip):
36 str ip, [sp, #-4]!
40 ldr ip, [sp, #+0]
44 str ip, [fp, #-12]
48 add sp, sp, #4
Now:
36 str ip, [fp, #-12]
Review URL: http://codereview.chromium.org/267118
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3076 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
fast code generator is optimized for compilation time and code size.
Currently it is only implemented on IA32. It is potentially triggered
for any code in the global scope (including code eval'd in the global
scope). It performs a syntactic check and chooses to compile in fast
mode if the AST contains only supported constructs and matches some
other constraints.
Initially supported constructs are
* ExpressionStatement,
* ReturnStatement,
* VariableProxy (variable references) to parameters and
stack-allocated locals,
* Assignment with lhs a parameter or stack-allocated local, and
* Literal
This allows compilation of literals at the top level and not much
else.
All intermediate values are allocated to temporaries and the stack is
used for all temporaries. The extra memory traffic is a known issue.
The code generated for 'true' is:
0 push ebp
1 mov ebp,esp
3 push esi
4 push edi
5 push 0xf5cca135 ;; object: 0xf5cca135 <undefined>
10 cmp esp,[0x8277efc]
16 jnc 27 (0xf5cbbb1b)
22 call 0xf5cac960 ;; code: STUB, StackCheck, minor: 0
27 push 0xf5cca161 ;; object: 0xf5cca161 <true>
32 mov eax,[esp]
35 mov [ebp+0xf4],eax
38 pop eax
39 mov eax,[ebp+0xf4]
42 mov esp,ebp ;; js return
44 pop ebp
45 ret 0x4
48 mov eax,0xf5cca135 ;; object: 0xf5cca135 <undefined>
53 mov esp,ebp ;; js return
55 pop ebp
56 ret 0x4
Review URL: http://codereview.chromium.org/273050
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3067 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
while, and for loops.
Previously they were distinguished by a type field, which required
runtime asserts to avoid invalid nodes (since not all loop types have
the same internal structure). Now they C++ type system is used to
require well-formed loop ASTs.
Because they do not share compilation code, we had very large
functions in the code generators that merely did a runtime dispatch to
a specific implementation based on the type.
Review URL: http://codereview.chromium.org/269049
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3048 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
The main piece of this change was to add support for break on return for ARM. On ARM the normal js function return consist of the following code sequence.
mov sp, fp
ldmia sp!, {fp, lr}
add sp, sp, #4
bx lr
to a call to the debug break return entry code using the following code sequence
mov lr, pc
ldr pc, [pc, #-4]
<debug break return entry code entry point address>
bktp 0
The values of Assembler::kPatchReturnSequenceLength and Assembler::kPatchReturnSequenceLength are somewhat misleading, but they fit the current use in the debugger. Also Assembler::kPatchReturnSequenceLength is used in the IC code as well (for something which is not related to return sequences at all). I will change that in a separate changelist.
For the debugger to work also added recording of the return sequence in the relocation info and handling of source position recording when a function ends with a return statement.
Used the constant kInstrSize instead of sizeof(Instr).
Passes all debugger tests on both simulator and hardware (only release mode tested on hardware).
Review URL: http://codereview.chromium.org/199075
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2879 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Objects which require an additional fixed array to be allocated now have this allocated in generated code as well. Added allocation flags to the macro assembler new space allocation routines.
Changed the ia32 and x64 macro assemblers to take allocation flags to the allocation routines instead of boolean flag.
Review URL: http://codereview.chromium.org/201015
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2832 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This change moves the allocation of new objects into generated code. The allocation will bail out into the runtime system if the number of properties to allocate for the object exceeds the number of in-object properties.
Review URL: http://codereview.chromium.org/175045
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2797 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
For objects which only have simple assignments of the form this.x = ...; a
specialized constructor stub is now generated. This generated code allocates the
object and fills in the initial properties directly. If this fails for some
reason code continues in the generic constructor stub which in turn might pass
control to the runtime system.
Added counter to see how many objects are constructed using a specialized stub.
The specialized stub is only implemented for ia32 architecture in this change.
For x64 and ARM the generic construct stub is used.
This is change is identical to http://codereview.chromium.org/174392 (committed in r2753 and reverted in r2754) except that a few parts have already been committed from http://codereview.chromium.org/173469 (committed in r2762).
Review URL: http://codereview.chromium.org/173470
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2764 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
For objects which only have simple assignments of the form this.x = ...; a specialized constructor stub is now generated. This generated code allocates the object and fills in the initial properties directly. If this fails for some reason code continues in the generic constructor stub which in turn might pass control to the runtime system.
Added counter to see how many objects are constructed using a specialized stub.
The specialized stub is only implemented for ia32 architecture in this change. For x64 and ARM the generic construct stub is used.
Review URL: http://codereview.chromium.org/174392
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2753 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
they do nothing. The frame is currently always spilled, so they were
not doing anything useful.
The call sites have been left alone to mark where spills will
eventually be needed if we begin doing register allocation on ARM.
Review URL: http://codereview.chromium.org/164136
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2644 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
fast-mode compiler.
1. We avoid generating a useless temporary for assignments with
nontrivial right-hand sides. Instead of translating id = expr into:
...
tmp = <last expr instruction>
id = tmp
we generate directly
...
id = <last expr instruction>
by passing a data destination ('hint') down the AST. The semantics is
to use the destination as a result location if a temp is needed. It
may be ignored. NULL indicates I don't care and you should generate a
temp.
2. We correctly handle assignments as subexpressions. When building
the CFG for an expression we accumulate the assigned variables and we
emit a move to a fresh temporary if a value in a variable is in
jeopardy of being overwritten.
Review URL: http://codereview.chromium.org/165056
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2643 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
generated in one-pass from the source AST, code is generated from the
CFG. Enabled by the flag --multipass and disabled by default.
Rudimentary and currently only supports literal expressions and return
statements. There are some other known limitations (e.g., missing
support for tracing).
Review URL: http://codereview.chromium.org/159695
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2596 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
We did not handle the case where the left-hand-side expression was
fully compiled to control flow. There were also some assertions for
unary and binary expressions that crashed debug builds when the
expression was fully compiled to control flow.
Regression test added.
Review URL: http://codereview.chromium.org/160006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2524 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
* Fast runtime calls for div and mod.
* Fix assembly and disassembly of multiply instructions.
* Strength reduce and inline multiplications to shift-add.
* Strength reduce and inline mod by power of 2.
* Strength reduce mod by other small integers to mul.
* Strength reduce div by 2 and 3.
Review URL: http://codereview.chromium.org/155047
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2355 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Remove the check for deleted properties in the global load inline cache if the property is known to be read only.
Propegate the in loop flag for the global call inline cache.
Changed the propagation of the code flags in the call stub compiler to compute these the same way for all types of call stubs and assert that the flags for the generated code is the same as those used for the cache lookup.
Addressed a few comments from previous review in test-api.cc.
Review URL: http://codereview.chromium.org/150101
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2308 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
* Fix incorrect signedness in disassembly of umull/mull on ARM.
* Fix incorrect register order in disassembly of umull/mull.
* Fix incorrect assembly of umull on ARM.
* Remove retroactively obsoleted restriction on choice of
registers in mul instructions on ARM.
Review URL: http://codereview.chromium.org/150002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2292 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Marsaglia's multiply-with-carry instead of mixing the
bits obtained from calling the system random() twice.
This seems to be a bit faster and gives a better
distribution than the system random() in particular on
Windows.
Review URL: http://codereview.chromium.org/126113
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2159 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
* Identify heap numbers that contain non-Smi int32s and do bit
ops on them without calling the fp hardware or emulation.
* Identify results that are non-Smi int32s and write them into
heap numbers without calling the fp hardware or emulation.
* Do unary minus on heap numbers without going into the runtime
system.
* On add, sub and mul if we have both Smi and heapnumber inputs
to the same operation then convert the Smi to a double and do
the op without going into runtime system. This also applies
if we have two Smi inputs but the result is not Smi.
Review URL: http://codereview.chromium.org/119241
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2131 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
instructions. The intention is that the snapshots generated
by the simulator should be usable on the hardware. Instead of
swi instructions we generate a branch to a swi instruction that
is not part of the snapshot. The call/jump is patched up in
the same way as other external references when the snapshot
is deserialized. This only works for EABI targets: on old ABI
targets we still emit some instructions not supported by the
simulator (fp coprocessor instructions).
Review URL: http://codereview.chromium.org/119036
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2127 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
deferred code snippets are highly stylized. They always make a call
to a stub or the runtime and then return. This change takes advantage
of that.
Creating a deferred code object now captures a snapshot of the
registers in the virtual frame. The registers are automatically saved
on entry to the deferred code and restored on exit.
The clients of deferred code must ensure that there is no change to
the registers in the virtual frame (eg, by allocating which can cause
spilling) or to the stack pointer. That is currently the case.
As a separate change, I will add either code to verify this constraint
or else code to forbid any frame effect.
The deferred code itself does not use the virtual frame or register
allocator (or even the code generator). It is raw macro assembler
code.
Review URL: http://codereview.chromium.org/118226
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2112 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
submitted in revisions 2093, 2094, 2099, and 2106.
There's no evidence that supports that these changes
should be the cause of the unexplained performance
regressions on the intl2 and DHTML page cyclers.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2109 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
encoding the values in one word and by using an indirection table for
handles.
This reduces compilation time by roughly 10% and we should be able to make the slow case equality checking of frame elements faster as well.
Review URL: http://codereview.chromium.org/115347
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1949 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
stack pointer to change by more than one in a corner case. If we push
a constant smi larger than 16 bits, we push it via a temporary
register. Allocating the temporary can cause a register to be spilled
from the frame somewhere above the stack pointer.
As a fix, do not use pushes to materialize ranges of elements of size
larger than one.
Review URL: http://codereview.chromium.org/92121
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1785 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
