Add a feedback vector slot for computed property names in object
and class literals. Introduce new slot kind for storing
computed property names.
Change StaDataPropertyInLiteral to use the accumulator (again), so
we don't exceed Bytecodes::kMaxOperands.
We assume that most computed property names are
symbols. Therefore we should see performance
improvements, even if we deal with monomorphic ICs only.
This CL only collects feedback but does not use
it in Reduce() yet.
BUG=v8:5624
Review-Url: https://codereview.chromium.org/2587393006
Cr-Commit-Position: refs/heads/master@{#42082}
port 72c370767226cf573d316655b1d3e3d3d699cc9b(r42034)
original commit message:
Currently the CompileLazy builtin checks the SFI expliciltly for FCG code. This means
if the SFI has bytecode we have to go through to the runtime to install the
interpreter entry trampoline into the JSFunction object.
Modify the builtin to always put the SFI code object into the JSFunction unless it's
the lazy compile stub on the SFI as well.
BUG=
Review-Url: https://codereview.chromium.org/2614533003
Cr-Commit-Position: refs/heads/master@{#42081}
port d3ed71ed87426046afe8938592588eeaa92ec7a9(r42021)
original commit message:
Don't fallback to the %StrictEqual / %Equal runtime functions for the
generic CompareIC slow path, but use the (new) StrictEqual and Equal
builtins instead. This avoids a performance cliff when mixing input
types for strict equality sites.
BUG=
Review-Url: https://codereview.chromium.org/2612693002
Cr-Commit-Position: refs/heads/master@{#42080}
For benchmarks with multiple measurements (e.g. runtime, compile
time, generated code size), in case of error, we'd get only the
last portion of the name ("Compile"), which is ambiguous and makes
debugging dificult. This change includes the full name, e.g.
"Unity/Physics/Compile".
BUG=
Review-Url: https://codereview.chromium.org/2615623003
Cr-Commit-Position: refs/heads/master@{#42078}
Add a more low level BranchIfFastPath to take the native_context and
promise_fun as args and change the existing one to use this.
BUG=v8:5343
Review-Url: https://codereview.chromium.org/2592933004
Cr-Commit-Position: refs/heads/master@{#42075}
This moves the initialization of [[HomeObject]] for constructors from
the %DefineClass runtime function into the bytecode generator, and
makes it conditional (resolving an old TODO). As part of this refactor,
avoid a load of "prototype" by returning the class prototype from
%DefineClass.
This is one of many steps in moving more of class definition into
bytecode.
R=rmcilroy@chromium.org
Review-Url: https://codereview.chromium.org/2610683003
Cr-Commit-Position: refs/heads/master@{#42072}
Remove a bunch of unnecessary v8:: prefixes in include/v8.h.
Some are still necessary for disambiguation purposes (e.g., between
bool Value() and v8::Value, or between v8::Isolate and
v8::internal::Isolate), but many aren't.
BUG=None
Review-Url: https://codereview.chromium.org/2605103003
Cr-Commit-Position: refs/heads/master@{#42071}
This removes all the promise allocation related methods from the CSA
and moves them PromiseBuiltinsAssembler with some edits.
BUG=v8:5343
Review-Url: https://codereview.chromium.org/2604273003
Cr-Commit-Position: refs/heads/master@{#42070}
Recognize the special
NumberFloor(NumberDivide(lhs, rhs))
subgraph in TypedOptimization, where both lhs and rhs are in the
Unsigned32 range, and the result is a PlainNumber, and replace the
NumberFloor with a NumberToUint32 truncation.
This could be done in a cleaner way if we have a dedicated type for all
double values in the Unsigned32 range, but that would complicate the
type system quite a bit.
R=jarin@chromium.org
BUG=v8:5267
Review-Url: https://codereview.chromium.org/2614663002
Cr-Commit-Position: refs/heads/master@{#42063}
mjsunit/big-object-literal will exhaust the stack in debug builds,
due to different compiler settings. It will work in optdebug.
This disables for both. Adding an 'optdebug' test to .status files
is easy, but I don't want to contribute to the 'mode x variant explosion'.
R=jochen@chromium.orgCC=franzih@chromium.org, machenbach@chromium.org
Review-Url: https://codereview.chromium.org/2609193002
Cr-Commit-Position: refs/heads/master@{#42061}
This CL simplifies the relation between the wasm graph builder, the
wasm decoder, and the wasm module they work on.
BUG=
Review-Url: https://codereview.chromium.org/2612643002
Cr-Commit-Position: refs/heads/master@{#42056}
-- Removes remaning debug from promise.js and moves it to c++
-- Changes debug_id to be a smi in PromiseReactionJobInfo and
PromiseResolveThenableJobInfo.
-- Changes debug_name to be a smi in PromiseReactionJobInfo and
PromiseResolveThenableJobInfo.
-- Adds PromiseDebugActionName and PromiseDebugActionType enums
-- Adds PromiseDebugActionNameToString and
PromiseDebugActionTypeToString helper methods
-- Changes variable `status` to be int in runtime functions.
-- Changes debug_id to start from 1, not 0 for easier bookkeeping.
BUG=v8:5343
Review-Url: https://codereview.chromium.org/2606093002
Cr-Commit-Position: refs/heads/master@{#42052}
These maps contain exactly the same information as the strict maps, so
this frees up a few pointers of native context space, gets rid of some
branches in FastNewClosure, and adds missing poisoned properties tests
for async functions.
BUG=v8:2355, v8:4483
R=adamk@chromium.org, bmeurer@chromium.org, littledan@chromium.org
Review-Url: https://codereview.chromium.org/2608333002
Cr-Commit-Position: refs/heads/master@{#42051}
Port 72c3707672
Original Commit Message:
Currently the CompileLazy builtin checks the SFI expliciltly for FCG code. This means
if the SFI has bytecode we have to go through to the runtime to install the
interpreter entry trampoline into the JSFunction object.
Modify the builtin to always put the SFI code object into the JSFunction unless it's
the lazy compile stub on the SFI as well.
R=rmcilroy@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=v8:4380
LOG=N
Review-Url: https://codereview.chromium.org/2602383003
Cr-Commit-Position: refs/heads/master@{#42046}
The previous patch for this bug (https://codereview.chromium.org/2599793002/) was wrong because it changed the behavior of isCreatedPhi() in an incompatible way. The actual source of the bug is that escape analysis propagates information along cycles without considering the previous analysis value. This fix makes sure that if a previous merge cleared a field, then it stays cleared.
R=bmeurer@chromium.org
BUG=chromium:670202
Review-Url: https://codereview.chromium.org/2610703002
Cr-Commit-Position: refs/heads/master@{#42045}
MarkCompactCollector::Prepare is too early in the rare case that we don't have a
marking deque yet. EnterFinalPause could then trigger
Heap::RegisterExternallyReferencedObject which rightfully fails because MC is
not properly set up.
Note that the DCHECK also triggers, without actual consequences, for aborting
incremental marking.
BUG=468240
TBR=jochen@chromium.org
Review-Url: https://codereview.chromium.org/2611753002
Cr-Commit-Position: refs/heads/master@{#42044}
Tracks background compilation of Ignition in a separate bucket from main thread
compilation. Also add some more compilation buckets for functions which can take a
significant proportion of compilation.
BUG=v8:5203,v8:5215
Review-Url: https://codereview.chromium.org/2577263002
Cr-Original-Commit-Position: refs/heads/master@{#42026}
Committed: b0e9116d59
Review-Url: https://codereview.chromium.org/2577263002
Cr-Commit-Position: refs/heads/master@{#42042}