Reason for revert:
[Sheriff] Breaks android runs. Will look later why.
Original issue's description:
> [tools] add --pretty switch to run_perf.py
>
> This CL improves running our internal benchmarks locally by adding the
> --pretty option to tools/run_perf.py. With the flag enabled we print
> the run-time of each benchmark directly and avoid the json output at
> the end.
>
> NOTRY=true
>
> Committed: https://crrev.com/83f69507ab1b9380b56758b747d4f3fabc849e49
> Cr-Commit-Position: refs/heads/master@{#33981}
TBR=cbruni@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/1696293002
Cr-Commit-Position: refs/heads/master@{#34017}
Fixes a bug in Ignition on Arm64 where lr gets trashed in StaContextSlot
which causes the stack walker to get confused and crash.
BUG=v8:4680
LOG=N
Review URL: https://codereview.chromium.org/1694263002
Cr-Commit-Position: refs/heads/master@{#34016}
This was actually only necessary for dealing with %_Arguments and
%_ArgumentsLength in Crankshaft, which have been removed recently,
so there's no need to keep this piece of awesomeness around any
longer.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1695193003
Cr-Commit-Position: refs/heads/master@{#34015}
This functionality is useful for stubs that need to walk the stack. The new
machine operator, LoadParentFramePointer dosn't force the currently compiling
method to have a frame in contrast to LoadFramePointer. Instead, it adapts
accordingly when frame elision is possible, making efficient stack walks
possible without incurring a performance penalty for small stubs that can
benefit from frame elision.
R=bmeurer@chromium.org
LOG=N
Review URL: https://codereview.chromium.org/1695313002
Cr-Commit-Position: refs/heads/master@{#34014}
port 4ff159bd28be36a39a1f8416cdf8fccafd3c2f95(r33880)
original commit message:
Add dedicated %LoadLookupSlot, %LoadLookupSlotInsideTypeof,
%LoadLookupSlotForCall, %StoreLookupSlot_Sloppy and
%StoreLookupSlot_Strict runtime entry points and use them
appropriately in the various compilers. This way we can
finally drop the machine operators from the JS graph level
completely in TurboFan.
Also drop the funky JSLoadDynamic operator from TurboFan,
which was by now just a small wrapper around the runtime
call to %LoadLookupSlot.
BUG=
Review URL: https://codereview.chromium.org/1694343002
Cr-Commit-Position: refs/heads/master@{#34012}
port d1c28849c77892ec74e58891aba44d5bfda8c0ba(r33873)
original commit message:
Moves InterpreterAssembler out of the compiler directory and into the
interpreter directory. Makes InterpreterAssembler as subclass of
CodeStubAssembler.
As part of this change, the special bytecode dispatch linkage type
is removed and instead we use a InterfaceDispatchDescriptor and
a normal CodeStub linkage type.
Removes a bunch of duplicated logic in InterpreterAssembler and
instead uses the CodeStubAssembler logic. Refactors Interpreter
with these changes.
Modifies CodeStubAssembler to add the extra operations required
by the Interpreter (extra call types, raw memory access and some extra
binary ops). Also adds the ability for subclasses to add extra
prologue and epilogue operations around calls, which is required
for the Interpreter.
BUG=
Review URL: https://codereview.chromium.org/1696263002
Cr-Commit-Position: refs/heads/master@{#34011}
There's no need to use a dedicated ChangeLoweringPhase before we go to
GenericLoweringPhase; instead that change lowering (which is really
simplified lowering by now) should run together with generic lowering.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1697133002
Cr-Commit-Position: refs/heads/master@{#34008}
Port 09d8453547
Original commit message:
The FastNewStrictArgumentsStub is very similar to the recently added
FastNewRestParameterStub, it's actually almost a copy of it, except that
it doesn't have the fast case we have for the empty rest parameter. This
patch improves strict arguments in TurboFan and fullcodegen by up to 10x
compared to the previous version.
Also introduce proper JSSloppyArgumentsObject and JSStrictArgumentsObject
for the in-object properties instead of having them as constants in the
Heap class.
Drive-by-fix: Use this stub and the FastNewRestParameterStub in the
interpreter to avoid the runtime call overhead for strict arguments
and rest parameter creation.
BUG=
Review URL: https://codereview.chromium.org/1698173002
Cr-Commit-Position: refs/heads/master@{#34003}
V8 tracks already most useful information, but lacks proper tracing scopes
that make it possible to distinguish certain events from each other.
- add trace-scope to track lazy-parsing due to optimization
- add trace-scope to track code optimization
BUG=
Review URL: https://codereview.chromium.org/1661883003
Cr-Commit-Position: refs/heads/master@{#34002}
The GraphTrimmer should not ever see a dead node, except for the roots
that are explicitly fed into it. To defend against this, turn the
condition into a DCHECK.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1698883003
Cr-Commit-Position: refs/heads/master@{#34001}
Eventually compiler.h might need to depend on the Pipeline to manage its
life cycle for concurrent recompilation, so we should not have the
cyclic include dependency here.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1698113002
Cr-Commit-Position: refs/heads/master@{#34000}
Harvesting maps from the stub cache for megamorphic ICs is both slow
(linear in the size of the stub cache) and imprecise (as it finds all
maps that have a cached handler for the given property name).
In the canonical megamorphic situation, this type feedback is useless
anyway. The interesting case is when we can filter it down to a single
map; however in these cases it is often possible to derive this map
just by looking at the HGraph, which is both faster and more reliable.
Review URL: https://codereview.chromium.org/1669213003
Cr-Commit-Position: refs/heads/master@{#33998}
Now the tool produces a far more readable output format, which bears a
lot of resemblance to YAML. In fact, the output should be machine
parseable as such, one document per testcase. However, the output format
may be subject to changes in future, so don't rely on this property.
In general, the output format has been optimized for producing a meaningful
textual diff, while keeping a decent readability as well. Therefore, not
everything is as compact as it could be, e.g. for an empty const pool we get:
constant pool: [
]
instead of:
constant pool: []
Also, trailing commas are always inserted in lists.
Additionally, now the tool accepts its output format as input. When
operating in this mode, all the snippets are extracted, processed and
the output is then emitted as usual. If nothing has changed, the output
should match the input. This is very useful for catching bugs in the
bytecode generation by running a textual diff against a known-good file.
The core (namely bytecode-expectations.cc) has been extracted from the
original cc file, which provides the utility as usual. The definitions
in the matching header of the library have been moved into the
v8::internal::interpreter namespace.
The library exposes a class ExpectationPrinter, with a method
PrintExpectation, which takes a test snippet as input, and writes the
formatted expectation to the supplied stream. One might then use a
std::stringstream to retrieve the results as a string and run it through
a diff utility.
BUG=v8:4280
LOG=N
Review URL: https://codereview.chromium.org/1688383003
Cr-Commit-Position: refs/heads/master@{#33997}
for the special case where the same register is used as both left and
right input.
Review URL: https://codereview.chromium.org/1695283002
Cr-Commit-Position: refs/heads/master@{#33996}
Passing floating point params to/from C has never quite worked correctly,
but we've never enforced the restriction early in the CallDescriptor
creation process because of unittests. Fix unittests to make their own
simple call descriptors and not rely on the C ones.
R=bmeurer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1701593003
Cr-Commit-Position: refs/heads/master@{#33993}
Properly type String.prototype.concat, String.prototype.charCodeAt,
and String.prototype.toLowerCase/toUpperCase in TurboFan. Also assign
better type to %_StringCharFromCode.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1701673002
Cr-Commit-Position: refs/heads/master@{#33991}
This is to enable deduplicating performance tests. We'll
create a hash of all relevant files and send it to perf bots
alongside the other swarming hashes (follow up on infra
side).
This will not actually run on swarming yet, but could at
some later point.
This splits off the cctest executable from other verification
test files, as those are not needed in performance tests.
BUG=chromium:535160
LOG=n
Review URL: https://codereview.chromium.org/1695243002
Cr-Commit-Position: refs/heads/master@{#33989}
Improve instruction selector for mask and shift operations by using cheaper
instructions where possible, in preference to UBFX.
Reverted because it was suspected of causing a couple of flaky tests to fail,
but investigation suggests this is unlikely.
Original review: https://codereview.chromium.org/1677023002
BUG=
Review URL: https://codereview.chromium.org/1684073006
Cr-Commit-Position: refs/heads/master@{#33988}
Turn the fast case of ArgumentsAccessStub into a new stub
FastNewSloppyArgumentsStub, which is similar to the existing
FastNewStrictArgumentsStub, although not polished yet, and the slow
case always went to the runtime anyway, so we can just directly emit
a runtime call there.
R=mstarzinger@chromium.org
Committed: https://crrev.com/55b0b4f6d572531eec00ab6ebd8f6feb7c584e04
Cr-Commit-Position: refs/heads/master@{#33973}
Review URL: https://codereview.chromium.org/1695633003
Cr-Commit-Position: refs/heads/master@{#33986}
port 5de27c343bbf898ca87246caa1e83e533ec44561(r33865)
original commit message:
Calls use registers for target, new_target and argument count.
We don't always respect argument count. It didn't bite us in the past
because the code paths where we clobbered it never used it, though
in future it could be an issue.
BUG=
Review URL: https://codereview.chromium.org/1698823002
Cr-Commit-Position: refs/heads/master@{#33984}
This CL improves running our internal benchmarks locally by adding the
--pretty option to tools/run_perf.py. With the flag enabled we print
the run-time of each benchmark directly and avoid the json output at
the end.
NOTRY=true
Review URL: https://codereview.chromium.org/1681283004
Cr-Commit-Position: refs/heads/master@{#33981}
Add a section identifier for declaring a start function as an index into
the function table. (This could also be done as a decl flag on the
function, but don't feel strongly here, since we probably want to redo
this when adding an import/export section.)
The start function must accept no parameters. Its return value is
currently ignored.
R=binji@chromium.org,bradnelson@chromium.org
BUG=chromium:575167
LOG=Y
Review URL: https://codereview.chromium.org/1692173002
Cr-Commit-Position: refs/heads/master@{#33978}
Adds support for ES6 super keyword and performing loads, stores, and
calls to super class members.
Implements SetHomeObject and enables ThisFunctionVariable.
BUG=v8:4280,v8:4682
LOG=N
Review URL: https://codereview.chromium.org/1689573004
Cr-Commit-Position: refs/heads/master@{#33977}
This is mostly preparation for allowing the function closure to be materialized.
As a drive-by fix, I have added ignition source position support to the frame inspector (this fixed some ignition test failures).
Review URL: https://codereview.chromium.org/1698743002
Cr-Commit-Position: refs/heads/master@{#33975}
Initially we were unable to address certain stack slots in the callee
part of the frame, including the function marker, therefore we had to
hack a reload of the function register into the OSR prologue. Now that
we are able to address all stack slots, we no longer need this hack.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1666073002
Cr-Commit-Position: refs/heads/master@{#33974}
Turn the fast case of ArgumentsAccessStub into a new stub
FastNewSloppyArgumentsStub, which is similar to the existing
FastNewStrictArgumentsStub, although not polished yet, and the slow
case always went to the runtime anyway, so we can just directly emit
a runtime call there.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/1695633003
Cr-Commit-Position: refs/heads/master@{#33973}
This adds initial support for inline allocation of object and array
literals to the JSCreateLowering pass. It's basically identical to
what Crankshaft does.
This also unstages the TurboFan escape analysis, as the lowering seems
to trigger a bunch of bugs in it; those bugs will be fixed separately,
and we will re-enable escape analysis afterwards.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1698783002
Cr-Commit-Position: refs/heads/master@{#33972}
port 3ef573e9f127345cd9d04d7f9f5e51bf169ae103(r33809)
original commit message:
Replace the somewhat awkward RestParamAccessStub, which would always
call into the runtime anyway with a proper FastNewRestParameterStub,
which is basically based on the code that was already there for strict
arguments object materialization. But for rest parameters we could
optimize even further (leading to 8-10x improvements for functions with
rest parameters), by fixing the internal formal parameter count:
Every SharedFunctionInfo has a formal_parameter_count field, which
specifies the number of formal parameters, and is used to decide whether
we need to create an arguments adaptor frame when calling a function
(i.e. if there's a mismatch between the actual and expected parameters).
Previously the formal_parameter_count included the rest parameter, which
was sort of unfortunate, as that meant that calling a function with only
the non-rest parameters still required an arguments adaptor (plus some
other oddities). Now with this CL we fix, so that we do no longer
include the rest parameter in that count. Thereby checking for rest
parameters is very efficient, as we only need to check whether there is
an arguments adaptor frame, and if not create an empty array, otherwise
check whether the arguments adaptor frame has more parameters than
specified by the formal_parameter_count.
The FastNewRestParameterStub is written in a way that it can be directly
used by Ignition as well, and with some tweaks to the TurboFan backends
and the CodeStubAssembler, we should be able to rewrite it as
TurboFanCodeStub in the near future.
Drive-by-fix: Refactor and unify the CreateArgumentsType which was
different in TurboFan and Ignition; now we have a single enum class
which is used in both TurboFan and Ignition.
BUG=
Review URL: https://codereview.chromium.org/1696063002
Cr-Commit-Position: refs/heads/master@{#33971}
port cfbd25617cfb8177bbb6377280e23ec356eb2373(r33857)
original commit message:
Preparing the young generation for (real) non-contiguous backing memory, this
change removes object masks that are used to compute containment in semi and new
space. The masks are replaced by lookups for object tags and page headers, where
possible.
Details:
- Use the fast checks (page header lookups) for containment in regular code.
- Use the slow version that masks out the page start adress and iterates all
pages of a space for debugging/verification.
- The slow version works for off-heap/unmapped memory.
- Encapsulate all checks for the old->new barrier in Heap::RecordWrite().
BUG=
Review URL: https://codereview.chromium.org/1698803002
Cr-Commit-Position: refs/heads/master@{#33970}
