This CL is a nightmare! For the utterly irrelevant edge case of a sloppy function with non-simple parameters and a call to direct eval, like here,
let x = 1;
function f(g = () => x) {
var y
eval("var x = 2")
return g() + x // f() = 3
}
we have to do all of the following, on top of the declaration block ("varblock") contexts we already introduce around the body:
- Introduce the ability for varblock contexts to have both a ScopeInfo and an extension object (e.g., the body varblock in the example will contain both a static var y and a dynamic var x). No other scope needs that. Since there are no context slots left, a special new struct is introduced that pairs up scope info and extension object.
- When declaring lookup slots in the runtime, this new struct is allocated in the case where an extension object has to be added to a block scope (at which point the block's extension slot still contains a plain ScopeInfo).
- While at it, introduce some abstraction to access context extension slots in a more controlled manner, in order to keep special-casing to a minimum.
- Make sure that even empty varblock contexts do not get optimised away when they contain a sloppy eval, so that they can host the potential extension object.
- Extend dynamic search for declaration contexts (used by sloppy direct eval) to recognize varblock contexts.
- In the parser, if a function has a sloppy direct eval, introduce an additional varblock scope around each non-simple (desugared) parameter, as required by the spec to contain possible dynamic var bindings.
- In the pattern rewriter, add the ability to hoist the named variables the pattern declares to an outer scope. That is required because the actual destructuring has to be evaluated inside the protecting varblock scope, but the bindings that the desugaring introduces are in the outer scope.
- ScopeInfos need to save the information whether a block is a varblock, to make sloppy eval calls work correctly that deserialise them as part of the scope chain.
- Add the ability to materialize block scopes with extension objects in the debugger. Likewise, enable setting extension variables in block scopes via the debugger interface.
- While at it, refactor and unify some respective code in the debugger.
Sorry, this CL is large. I could try to split it up, but everything is rather entangled.
@mstarzinger: Please review the changes to contexts.
@yangguo: Please have a look at the debugger stuff.
R=littledan@chromium.org, mstarzinger@chromium.org, yangguo@chromium.org
BUG=v8:811,v8:2160
LOG=N
Review URL: https://codereview.chromium.org/1292753007
Cr-Commit-Position: refs/heads/master@{#30295}
- Introduce a proper bit for SIMD primitive values.
- Introduce constructors for individual SIMD types. These are currently just classes, which seems good enough for now, given that we always have exactly one global map per SIMD type.
The only problem with using class types for SIMD is that a SIMD constant won't be a subtype of its specific type, only of the general SIMD type. But until we actually introduce SIMD constants into the compiler that shouldn't matter.
R=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1303863002
Cr-Commit-Position: refs/heads/master@{#30294}
port cbbaf9ea6a (r30224).
original commit message:
[turbofan] Unify referencing of stack slots
Previously, it was not possible to specify StackSlotOperands for all
slots in both the caller and callee stacks. Specifically, the region
of the callee's stack including the saved return address, frame
pointer, function pointer and context pointer could not be addressed
by the register allocator/gap resolver.
In preparation for better tail call support, which will use the gap
resolver to reconcile outgoing parameters, this change makes it
possible to address all slots on the stack, because slots in the
previously inaccessible dead zone may become parameter slots for
outgoing tail calls. All caller stack slots are accessible as they
were before, with slot -1 corresponding to the last stack
parameter. Stack slot indices >= 0 access the callee stack, with slot
0 corresponding to the callee's saved return address, 1 corresponding
to the saved frame pointer, 2 corresponding to the current function
context, 3 corresponding to the frame marker/JSFunction, and slots 4
and above corresponding to spill slots.
The following changes were specifically needed:
* Frame has been changed to explicitly manage three areas of the
callee frame, the fixed header, the spill slot area, and the
callee-saved register area.
* Conversions from stack slot indices to fp offsets all now go through
a common bottleneck: OptimizedFrame::StackSlotOffsetRelativeToFp
* The generation of deoptimization translation tables has been changed
to support the new stack slot indexing scheme. Crankshaft, which
doesn't support the new slot numbering in its register allocator,
must adapt the indexes when creating translation tables.
* Callee-saved parameters are now kept below spill slots, not above,
to support saving only the optimal set of used registers, which is
only known after register allocation is finished and spill slots
have been allocated.
BUG=
Review URL: https://codereview.chromium.org/1293103003
Cr-Commit-Position: refs/heads/master@{#30292}
port 00df60d1c6 (r30219).
original commit message:
Makes the following modifications to the interpreter builtins and
InterpreterAssembler:
- Adds an accumulator register and initializes it to undefined()
- Adds a register file pointer register and use it instead of FramePointer to
access registers
- Modifies builtin to support functions with 0 regiters in the register file
- Modifies builtin to Call rather than TailCall to first bytecode handler.
BUG=
Review URL: https://codereview.chromium.org/1304593002
Cr-Commit-Position: refs/heads/master@{#30289}
port f4c079d450 (r30107).
This is the appendix of 458dfe3b943edb3238917edfe9e2dde326cd1adb which misses
one modified file.
original commit message:
There's no need to have one InstanceType per SIMD primitive type (this
will not scale long-term). Also reduce the amount of code duplication
and make it more robust wrt adding new SIMD types.
BUG=
Review URL: https://codereview.chromium.org/1304963003
Cr-Commit-Position: refs/heads/master@{#30288}
Not all parenthesized AssignmentExpressions whose components are valid
binding patterns are valid arrow function formal parameters. In
particular (a,b,c)() is not valid, and in general the existing code
wasn't catching the tail productions of ConditionalExpression,
BinaryExpression, PostfixExpression, LeftHandSideExpression,
and MemberExpression.
Thanks to Adrian Perez for the test case.
BUG=v8:4211
LOG=Y
R=rossberg@chromium.org
Review URL: https://codereview.chromium.org/1306583002
Cr-Commit-Position: refs/heads/master@{#30286}
We can use the script type to determine that instead. Script of type
TYPE_NATIVE are considered builtins, TYPE_NORMAL are not. The only exception
to this rule is the empty function, for which the script is TYPE_NATIVE
(observable by the debugger), but should be stringified to "function () {}"
instead of "function () { [native code] }". For this, I introduce a
hide_source flag on the script object.
We also use IsBuiltin and IsSubjectToDebugging interchangeably. For debugger,
we now use the latter, hiding the detail that only non-builtins are debuggable.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/1292283004
Cr-Commit-Position: refs/heads/master@{#30285}
GC flags are now part of the {Heap} and should be respected by all
sub-components.
Also add a infrastructure to write tests accessing private methods.
Review URL: https://codereview.chromium.org/1301183002
Cr-Commit-Position: refs/heads/master@{#30281}
We currently have several ways to share symbols that are used in
both native scripts and the runtime. This change unifies this.
We do not use the symbols registry since we don't need the
registry any longer after bootstrapping, but the registry stays
alive afterwards.
R=mlippautz@chromium.org, rossberg@chromium.org
Review URL: https://codereview.chromium.org/1293493004
Cr-Commit-Position: refs/heads/master@{#30280}
This prevents leakage of the memory-reducer.h declarations inside of the
heap and prevents it from being exposed to the world. Protects private
state from being inadvertently mocked with.
R=mlippautz@chromium.org
Review URL: https://codereview.chromium.org/1288913003
Cr-Commit-Position: refs/heads/master@{#30277}
This prevents leakage of the gc-tracer.h declarations inside of the
heap and prevents it from being exposed to the world. Protects private
state from being inadvertently mocked with.
R=mlippautz@chromium.org
Review URL: https://codereview.chromium.org/1294763004
Cr-Commit-Position: refs/heads/master@{#30276}
Additionally, this CL moves a bit of code around to free up more memory before compaction starts.
BUG=
Review URL: https://codereview.chromium.org/1305733003
Cr-Commit-Position: refs/heads/master@{#30275}
Invalid slots in large objects can only occur when large objects became dead. These slots are filtered out already after marking.
BUG=
Review URL: https://codereview.chromium.org/1298183003
Cr-Commit-Position: refs/heads/master@{#30273}
This CL us a pure refactoring that makes an empty compilation unit
including just "snapshot.h" but not "objects-inl.h" compile without
warnings or errors. This is needed to further reduce the header
dependency tangle.
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/1287113010
Cr-Commit-Position: refs/heads/master@{#30270}
This make inclusion of unicode-inl.h in object.h absolete. Now most
compilation units don't require that header. It also breaks a cycle
within declarations of the scanner.h header.
This tries to remove includes of "-inl.h" headers from normal ".h"
headers, thereby reducing the chance of any cyclic dependencies and
decreasing the average size of our compilation units.
Note that this change still leaves 3 violations of that rule in the
code, checked with the "tools/check-inline-includes.sh" tool.
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/1287893006
Cr-Commit-Position: refs/heads/master@{#30268}
FlushICache should be NOP for Denver with part numbers 0x0, 0x1 and 0x2 only.
Instruction cache needs to flushed for future versions of denver.
Review URL: https://codereview.chromium.org/1287173004
Cr-Commit-Position: refs/heads/master@{#30262}
Rolling v8/third_party/android_tools to 4238a28593b7e6178c95431f91ca8c24e45fa7eb
Rolling v8/tools/clang to 1c7f9147c834d78b36787f31ecfc5c47f3c98da8
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/1293873004
Cr-Commit-Position: refs/heads/master@{#30261}
Embedders would use these for features which must be able to be turned
off at runtime, despite being compiled into V8. They can be turned on
and off by the embedder using the --experimental_extras flag, e.g. via
v8::SetFlagsFromString.
R=yangguo@chromium.org, mlippautz@chromium.org, hpayer@chromium.org
BUG=chromium:507137
LOG=Y
Review URL: https://codereview.chromium.org/1284413002
Cr-Commit-Position: refs/heads/master@{#30260}
