- fix truthfulness of comments
- use InitializeFieldsWithFiller more consistently
- use unsigned comparisons for pointers
No change in functionality intended.
Bonus: improve JavaScriptFrame::Print() for an enhanced debugging experience:
- print PC of each frame
- print the function's source also for optimized frames
Review URL: https://codereview.chromium.org/1186823003
Cr-Commit-Position: refs/heads/master@{#29082}
The remaining uses need some non-mechanical work:
* non-standard-layout type, probably due to mixed access control
* extended field designators
Review URL: https://codereview.chromium.org/1173343006
Cr-Commit-Position: refs/heads/master@{#29071}
This patch implements %_IsTypedArray in fullcodegen, Hydrogen and
Turbofan in order to implement fast type checks to enable ES6
TypedArray features and semantics efficiently.
R=adamk,titzer
LOG=Y
BUG=v8:4085
Review URL: https://codereview.chromium.org/1183213002
Cr-Commit-Position: refs/heads/master@{#29033}
Up until now we can only inline based on JSFunction, because of the way
the deoptimization works. With this change we will be able to inline
based on the SharedFunctionInfo and materialize the JSFunction from a
literal or a stack slot when necessary.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1169103004
Cr-Commit-Position: refs/heads/master@{#28906}
- Introduce Assembler::DataAlign for table alignment in code object
- Fix several misuses of r8 (alias of the pool pointer register, pp)
- Fix calculation of pp in OSR/handler entry invocation
- Enable missing cases in deserializer
- Fix references to ool constant pools in comments.
R=rmcilroy@chromium.org, michael_dawson@ca.ibm.com
BUG=chromium:497180
LOG=N
Review URL: https://codereview.chromium.org/1155673005
Cr-Commit-Position: refs/heads/master@{#28873}
These values were computed by the parser and hence out of sync with any
visitor over the AST. Our AST visitor aborts visitation of statement
lists as soon as a jump statement has been reached. Now handler tables
are guaranteed to be dense and fully populated.
R=ishell@chromium.org
TEST=mjsunit/regress/regress-crbug-493290
BUG=chromium:493290
LOG=N
Review URL: https://codereview.chromium.org/1157213004
Cr-Commit-Position: refs/heads/master@{#28846}
Boolean "is_strong" parameters have begun to proliferate across areas where
strong mode semantics are different. This CL repurposes the existing
ObjectStrength enum as a replacement for them.
BUG=v8:3956
LOG=N
Review URL: https://codereview.chromium.org/1144183004
Cr-Commit-Position: refs/heads/master@{#28839}
This unifies methods Deoptimizer::DoTranslateCommand, Deotpimizer::DoTranslateObject and the arguments object materializer.
To unify these, we have to separate reading of the input frame from writing to the output frame because the argument materializer does not write to output frames.
Instead, we now deoptimize in following stages:
1. Read out the input frame/registers, decode them using the translations from the deoptimizer and store them in the deoptimizer (Deoptimizer::translated_state_). This is done in TranslatedState::Init.
2. Write out into the output frame buffer all the values that do not require allocation. We also remember references to the values that require materialization. As before, this is done in Deoptimizer::DoCompute*Frame method, but instead calling to DoTranslateCommand, we use the translated frame to obtain the values and write them to the output frames.
3. The platform specific code then sets up the output frames and calls into the deoptimization notification. This has not been changed at all.
4. Once the stack is setup, we handlify all the references in the saved translated values (TranslatedState::Prepare).
5. Finally, we materialize all the values we remembered in step (1) and write them to their frames on the stack (using the TranslatedValue::GetValue method).
BUG=
Review URL: https://codereview.chromium.org/1136223004
Cr-Commit-Position: refs/heads/master@{#28826}
Before this we had 3 super related lexical bindings that got injected
into method bodies: .home_object, .this_function, and new.target.
With this change we get rid of the .home_object one in favor of using
.this_function[home_object_symbol] which allows some simplifications
throughout the code base.
BUG=v8:3768
LOG=N
R=adamk@chromium.org, wingo@igalia.com
Review URL: https://codereview.chromium.org/1154103005
Cr-Commit-Position: refs/heads/master@{#28802}
Embed constant pools within their corresponding Code
objects.
This removes support for out-of-line constant pools in favor
of the new approach -- the main advantage being that it
eliminates the need to allocate and manage separate constant
pool array objects.
Currently supported on PPC and ARM. Enabled by default on
PPC only.
This yields a 6% improvment in Octane on PPC64.
R=bmeurer@chromium.org, rmcilroy@chromium.org, michael_dawson@ca.ibm.com
BUG=chromium:478811
LOG=Y
Review URL: https://codereview.chromium.org/1162993006
Cr-Commit-Position: refs/heads/master@{#28801}
This fixes a corner-case where arrow functions that require a context
allocate none, because there are no additional slots allocated. Note
that this didn't happen with true function scopes because they always
had at least the receiver slot.
The outcome was a context chain that no longer was in sync with the
scope chain, hence context slot loads were bogus. This is observable
using the DYNAMIC_LOCAL optimization in all compilers.
R=rossberg@chromium.org,wingo@igalia.com
TEST=mjsunit/harmony/regress/regress-4160
BUG=v8:4160
LOG=N
Review URL: https://codereview.chromium.org/1146063006
Cr-Commit-Position: refs/heads/master@{#28788}
Previously the %_DateField intrinsic would also check the object and
throw an exception if you happen to pass something that is not a valid
JSDate, which (a) violates our policy for instrinsics and (b) is hard to
optimize in TurboFan (even Crankshaft has a hard time, but there we will
never inline the relevant builtins, so it doesn't show up). The throwing
part is now a separate intrinsics %_ThrowIfNotADate that throws an
exception in full codegen and deoptimizes in Crankshaft, which means the
code for the current use cases is roughly the same (modulo some register
renamings/gap moves).
R=jkummerow@chromium.org
Review URL: https://codereview.chromium.org/1167813003
Cr-Commit-Position: refs/heads/master@{#28782}
This splits the SuperReference AST node into SuperPropertyReference and
SuperCallReference. The super call reference node consists of three
unresolved vars to this, new.target and this_function. These gets
declared when the right function is entered and if it is in use. The
variables gets assigned in FullCodeGenerator::Generate.
This is a revert of the revert 88b1c9170a
BUG=v8:3768
LOG=N
R=wingo@igalia.com, adamk@chromium.org
Review URL: https://codereview.chromium.org/1168513004
Cr-Commit-Position: refs/heads/master@{#28769}
It's necessary to do this in order to know how many type feedback vector slots
we should allocate for the object literal.
BUG=
Review URL: https://codereview.chromium.org/1165463007
Cr-Commit-Position: refs/heads/master@{#28749}
When compiling on a laptop I like to concatenate the small test files.
This makes a big difference to compile times. These changes make that
easier.
R=ulan@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1163803002
Cr-Commit-Position: refs/heads/master@{#28742}
This splits the SuperReference AST node into SuperPropertyReference and
SuperCallReference. The super call reference node consists of three
unresolved vars to this, new.target and this_function. These gets
declared when the right function is entered and if it is in use. The
variables gets assigned in FullCodeGenerator::Generate.
BUG=v8:3768
LOG=N
R=wingo@igalia.com, adamk@chromium.org
Review URL: https://codereview.chromium.org/1146863007
Cr-Commit-Position: refs/heads/master@{#28731}
In a nutshell: The FILTER_KEY builtin is gone, and was replaced by a
simple runtime call to ForInFilter, which does everything and is even
cheaper (because FILTER_KEY used to call into the runtime anyway).
And ForInFilter returns either the name or undefined, which makes it
possible to remove the control flow construction from the AstGraphBuilder,
and thereby make both the initialization and the per-loop code of for-in
optimizable later (in typed lowering).
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1160983004
Cr-Commit-Position: refs/heads/master@{#28711}
vmov immediate was always emitting movt for some immediates wether or
not the CPU supported ARMv7.
BUG=v8:4019
LOG=n
Review URL: https://codereview.chromium.org/1160093002
Cr-Commit-Position: refs/heads/master@{#28691}
The list of inlined functions is used in exactly two places - for live
edit and to prevent code flushing for inlined functions - and those are
fine with SharedFunctionInfo and don't require a closure.
This is one additional step towards inlining based on SharedFunctionInfo
instead of JSFunction.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1145893003
Cr-Commit-Position: refs/heads/master@{#28672}
Also adapt code generation to pass the slot to the
store/keyed-store ic. AST nodes ObjectLiteral, Assignment,
ForEach, Call and CountOperation now include one or more
feedback vector ic slot ids.
BUG=
Review URL: https://codereview.chromium.org/1161623002
Cr-Commit-Position: refs/heads/master@{#28659}
When we enter a method that needs access to the [[HomeObject]]
we allocate a local variable `.home_object` and assign it the
value from the [[HomeObject]] private symbol. Something along
the lines of:
method() {
var .home_object = %ThisFunction()[home_object_symbol];
...
}
BUG=v8:3867, v8:4031
LOG=N
Review URL: https://codereview.chromium.org/1135243004
Cr-Commit-Position: refs/heads/master@{#28644}
* Hash code is now just done with a private own symbol instead of the hidden string, which predates symbols.
* In the long run we should do all hidden properties this way and get rid of the
hidden magic 0-length string with the zero hash code. The advantages include
less complexity and being able to do things from JS in a natural way.
* Initially, the performance of weak set regressed, because it's a little harder
to do the lookup in C++. Instead of heroics in C++ to make things faster I
moved some functionality into JS and got the performance back. JS is supposed to be good at looking up named properties on objects.
* This also changes hash codes of Smis so that they are always Smis.
Performance figures are in the comments to the code review. Summary: Most of js-perf-test/Collections is neutral. Set and Map with object keys are 40-50% better. WeakMap is -5% and WeakSet is +9%. After the measurements, I fixed global proxies, which cost 1% on most tests and 5% on the weak ones :-(.
In the code review comments is a patch with an example of the heroics we could do in C++ to make lookup faster (I hope we don't have to do this. Instead of checking for the property, then doing a new lookup to insert it, we could do one lookup and handle the addition immediately). With the current benchmarks above this buys us nothing, but if we go back to doing more lookups in C++ instead of in stubs and JS then it's a win.
In a similar vein we could give the magic zero hash code to the hash code
symbol. Then when we look up the hash code we would sometimes see the table
with all the hidden properties. This dual use of the field for either the hash
code or the table with all hidden properties and the hash code is rather ugly,
and this CL gets rid of it. I'd be loath to bring it back. On the benchmarks quoted above it's slightly slower than moving the hash code lookup to JS like in this CL.
One worry is that the benchmark results above are more monomorphic than real
world code, so may be overstating the performance benefits of moving to JS. I
think this is part of a general issue we have with handling polymorphic code in
JS and any solutions there will benefit this solution, which boils down to
regular property access. Any improvement there will lift all boats.
R=adamk@chromium.org, verwaest@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1149863005
Cr-Commit-Position: refs/heads/master@{#28622}
The reason is that this information will be needed to compute the number of
vector ic slots done at numbering time.
BUG=
Review URL: https://codereview.chromium.org/1150323002
Cr-Commit-Position: refs/heads/master@{#28615}
Also introduce new interface descriptors for the trampoline and full
versions of those stubs.
Currently, the stubs aren't functional.
BUG=
Review URL: https://codereview.chromium.org/1149903005
Cr-Commit-Position: refs/heads/master@{#28597}
This allows you to put iterables into your array literals
and the will get spread into the array.
let x = [0, ...range(1, 3)]; // [0, 1, 2]
This is done by treating the array literal up to the first
spread element as usual, including using a boiler plate
array, and then appending the remaining expressions and rest
expressions.
BUG=v8:3018
LOG=N
Review URL: https://codereview.chromium.org/1125183008
Cr-Commit-Position: refs/heads/master@{#28534}
Alas, this involved quite a bit of copy-n-paste between the
architectures, but this is caused by the very convoluted
relationships, lifetimes and distribution of responsibilities. This
should really be cleaned up by moving code around and using STL maps,
but that's not really a priority right now.
Bonus: Fixed leaks in the ARM64 disassembler tests.
Review URL: https://codereview.chromium.org/1132943007
Cr-Commit-Position: refs/heads/master@{#28496}
Also check whether the arguments count is smaller than the number of
required parameters which is the same as the SharedFunctionInfo length.
BUG=v8:4102
LOG=N
R=rossberg@chromium.org
Review URL: https://codereview.chromium.org/1133933003
Cr-Commit-Position: refs/heads/master@{#28491}
... and the following two
"PPC: Resolve references to "this" the same way as normal variables"
"Remove Scope::scope_uses_this_ flag"
R=hablich@chromium.org
BUG=chromium:487289
LOG=N
Review URL: https://codereview.chromium.org/1134003003
Cr-Commit-Position: refs/heads/master@{#28395}
Implements the strong mode proposal's restrictions on implicit conversions
for the binary >, >=, <, and <= operators.
BUG=v8:3956
LOG=N
Review URL: https://codereview.chromium.org/1130283002
Cr-Commit-Position: refs/heads/master@{#28370}
HMaybeGrowElements moves the situation where you actually have to grow
into deferred code. This means crankshaft doesn't have to spill registers
just to make the bounds comparison to see if it'll need to grow or not.
It makes the growing case a bit more expensive, but reduces the cost of
the general case.
BUG=chromium:484025
LOG=N
Review URL: https://codereview.chromium.org/1124443004
Cr-Commit-Position: refs/heads/master@{#28359}
