port d6ee366d5c (r29834).
original commit message:
This is the initial round of optimizations for the
LoadGlobalViaContextStub and StoreGlobalViaContextStub, basically
turning them into platform code stubs to avoid the Crankshaft overhead
in the fast case, and making the runtime interface cheaper.
BUG=
Review URL: https://codereview.chromium.org/1258513003
Cr-Commit-Position: refs/heads/master@{#29839}
port 7877c4e0c7 (r29814).
original commit message:
Adds basic support for generation of interpreter bytecode handler code
snippets. The InterpreterAssembler class exposes a set of low level,
interpreter specific operations which can be used to build a Turbofan
graph. The Interpreter class generates a bytecode handler snippet for
each bytecode by assembling operations using an InterpreterAssembler.
Currently only two simple bytecodes are supported: LoadLiteral0 and Return.
BUG=
Review URL: https://codereview.chromium.org/1256783002
Cr-Commit-Position: refs/heads/master@{#29838}
port bc8041dc2b (r29811).
original commit message:
Previous to this CL, ICs used a slightly different code idiom
to get to C++ code from generated code than runtime intrinsics,
using an IC_Utility class that in essence provided exactly
the same functionality as Runtime::FunctionForId, but in its
own quirky way.
This CL unifies the two mechanisms, folding IC_Utility
away by making all IC entry points in C++ code, e.g. IC
miss handlers, full-fledged runtime intrinsics. This makes
it possible to eliminate a bunch of ad-hoc declarations and
adapters that the IC system had to needlessly re-invent.
As a bonus and the original reason for this yak-shave:
IC-related C++ runtime functions are now callable from
TurboFan.
BUG=
Review URL: https://codereview.chromium.org/1252903002
Cr-Commit-Position: refs/heads/master@{#29837}
port 3334b830a5 (r20813).
original commit message:
HydrogenCodeStubs consume stack arguments via descriptor.
All of this is controlled by the CallDescriptor. It's simply the case
that if you specify less registers than the function arity calls for,
the rest are assumed to be on the stack.
Bailout handlers accept these constant stack arguments too.
BUG=
Review URL: https://codereview.chromium.org/1258553002
Cr-Commit-Position: refs/heads/master@{#29836}
This is the initial round of optimizations for the
LoadGlobalViaContextStub and StoreGlobalViaContextStub, basically
turning them into platform code stubs to avoid the Crankshaft overhead
in the fast case, and making the runtime interface cheaper.
R=ishell@chromium.org
BUG=chromium:510694
LOG=n
Review URL: https://codereview.chromium.org/1238143002
Cr-Commit-Position: refs/heads/master@{#29834}
$toLength is slow, causing a 3.8%-8% regression in the Octane RegExp
benchmark. Reverting this patch brings it back up. To make this change,
we'll need a faster implementation fo $toLength.
BUG=chromium:513160
LOG=Y
R=adamk
Review URL: https://codereview.chromium.org/1243053005
Cr-Commit-Position: refs/heads/master@{#29830}
port 1f295980b7 (r29787).
original commit message:
This fixes a recent regression where the register holding the original
receiver was pushed onto the stack before the internal frame within the
CallStubInRecordCallTarget helper was created. That in turn confused
the stack walker when allocations in these stubs failed.
BUG=
Review URL: https://codereview.chromium.org/1247493004
Cr-Commit-Position: refs/heads/master@{#29828}
Adding CHECKED_OUT_VERSION in test/simdjs so that
the bots don't revert and redownload each time.
LOG=N
BUG=None
TEST=None
R=littledan@chromium.org,machenbach@chromium.org
Review URL: https://codereview.chromium.org/1242863003
Cr-Commit-Position: refs/heads/master@{#29826}
Port 1f295980b7
Original commit message:
This fixes a recent regression where the register holding the original
receiver was pushed onto the stack before the internal frame within the
CallStubInRecordCallTarget helper was created. That in turn confused
the stack walker when allocations in these stubs failed.
R=mstarzinger@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=chromium:512711
LOG=N
Review URL: https://codereview.chromium.org/1252493004
Cr-Commit-Position: refs/heads/master@{#29823}
Port 7877c4e0c7
Original commit message:
Adds basic support for generation of interpreter bytecode handler code
snippets. The InterpreterAssembler class exposes a set of low level,
interpreter specific operations which can be used to build a Turbofan
graph. The Interpreter class generates a bytecode handler snippet for
each bytecode by assembling operations using an InterpreterAssembler.
Currently only two simple bytecodes are supported: LoadLiteral0 and Return.
R=rmcilroy@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/1250723003
Cr-Commit-Position: refs/heads/master@{#29822}
Port bc8041dc2b
Original commit message:
Previous to this CL, ICs used a slightly different code idiom
to get to C++ code from generated code than runtime intrinsics,
using an IC_Utility class that in essence provided exactly
the same functionality as Runtime::FunctionForId, but in its
own quirky way.
This CL unifies the two mechanisms, folding IC_Utility
away by making all IC entry points in C++ code, e.g. IC
miss handlers, full-fledged runtime intrinsics. This makes
it possible to eliminate a bunch of ad-hoc declarations and
adapters that the IC system had to needlessly re-invent.
As a bonus and the original reason for this yak-shave:
IC-related C++ runtime functions are now callable from
TurboFan.
R=danno@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/1249433004
Cr-Commit-Position: refs/heads/master@{#29821}
This mostly removes dead code and obsolete special cases from the
RawMachineAssembler::MakeNode helper, that shouldn't be necessary
anymore.
R=rmcilroy@chromium.org
Review URL: https://codereview.chromium.org/1243253005
Cr-Commit-Position: refs/heads/master@{#29817}
According to the ECMA spec, a 24th hour is allowed if the minutes, seconds, and milliseconds are all zero (i.e. it's midnight). Previously, we parsed the date correctly, however, we failed to account in all checks for the possibility of a 24th hour. This CL changes the check to allow a 24th hour if it's exactly midnight.
BUG=chromium:174609
LOG=Y
Review URL: https://codereview.chromium.org/1240093005
Cr-Commit-Position: refs/heads/master@{#29816}
Adds basic support for generation of interpreter bytecode handler code
snippets. The InterpreterAssembler class exposes a set of low level,
interpreter specific operations which can be used to build a Turbofan
graph. The Interpreter class generates a bytecode handler snippet for
each bytecode by assembling operations using an InterpreterAssembler.
Currently only two simple bytecodes are supported: LoadLiteral0 and Return.
BUG=v8:4280
LOG=N
Review URL: https://codereview.chromium.org/1239793002
Cr-Commit-Position: refs/heads/master@{#29814}
All of this is controlled by the CallDescriptor. It's simply the case
that if you specify less registers than the function arity calls for,
the rest are assumed to be on the stack.
Bailout handlers accept these constant stack arguments too.
BUG=
Review URL: https://codereview.chromium.org/1250563004
Cr-Commit-Position: refs/heads/master@{#29813}
While at it, remove the notion of INTERNAL variables.
@caitp: Took some parts from your CL, since I was blocked on the temp scope bug.
R=mstarzinger@chromium.org
BUG=512574
LOG=N
Review URL: https://codereview.chromium.org/1250513004
Cr-Commit-Position: refs/heads/master@{#29812}
Previous to this CL, ICs used a slightly different code idiom
to get to C++ code from generated code than runtime intrinsics,
using an IC_Utility class that in essence provided exactly
the same functionality as Runtime::FunctionForId, but in its
own quirky way.
This CL unifies the two mechanisms, folding IC_Utility
away by making all IC entry points in C++ code, e.g. IC
miss handlers, full-fledged runtime intrinsics. This makes
it possible to eliminate a bunch of ad-hoc declarations and
adapters that the IC system had to needlessly re-invent.
As a bonus and the original reason for this yak-shave:
IC-related C++ runtime functions are now callable from
TurboFan.
Review URL: https://codereview.chromium.org/1248303002
Cr-Commit-Position: refs/heads/master@{#29811}
There are many test names in the v8 code base that prefix
others, which makes it hard to only run those tests.
BUG=chromium:511215
LOG=n
NOTRY=true
Review URL: https://codereview.chromium.org/1251363002
Cr-Commit-Position: refs/heads/master@{#29809}
In particular, rename FormalParameterParsingState and friends to FormalParameters etc.
This should not change any logic, but is a preparatory CL for a bunch of follow-up fixes and clean-ups.
R=mstarzinger@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1247443004
Cr-Commit-Position: refs/heads/master@{#29807}
Using the GraphBuilder base class forces each node creation to go
through a virtual function dispatch just for the sake of saving the
duplication of the NewNode helper methods. In total that added up to
saving minus (sic!) six lines of code.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/1252093002
Cr-Commit-Position: refs/heads/master@{#29799}
There is already non-throwing version FrameMirror.restart and RestartFrame in the livedit-debugger.js just adds a throwing wrapper around it.
Also NEEDS_STEP_IN_PROPERTY_NAME was removed, the client code can decide based on the stack_modified field if "step in" is required.
Chromium side was fixed in https://codereview.chromium.org/1249013002/
BUG=None
LOG=N
Review URL: https://codereview.chromium.org/1247363002
Cr-Commit-Position: refs/heads/master@{#29797}
In simulator data trace, DSLL did not print result and
BAL/BGEZAL omitted result from an instruction executed
in delay slot.
TEST=cctest/test-assembler-mips[64]
BUG=
Review URL: https://codereview.chromium.org/1245173002
Cr-Commit-Position: refs/heads/master@{#29796}
This preserves the context-independent entry in an optimized code map
across GCs when the code is considered young (i.e. less than 3 ages).
Note that any context-dependent entry for the same code will still be
flushed immediately when the respective context dies, hence context
lifetime is not increased.
R=hpayer@chromium.org
Review URL: https://codereview.chromium.org/1252463002
Cr-Commit-Position: refs/heads/master@{#29790}
