Currently, if the size of two cmp or test operands is a byte or a word, we sign-extend or zero-extend each of them into a 32-bit register before doing the comparison, even when the conditions for the use of a memory operand are met.
This CL makes it possible to load only one of them into a register and address the other as a memory operand.
Meanwhile, comparisons between Uint8 values in the string relational comparison stubs are done with Uint32LessThan (previously we were always zero-extending the byte to a 32-bit value, so signed comparison was alright).
Found that Assembler::arithmetic_op_8(byte, Register, const Operand&) wasn't taking the Operand's rex_ field into account, so I fixed that too.
BUG=
Review URL: https://codereview.chromium.org/1780193003
Cr-Commit-Position: refs/heads/master@{#34862}
(The goal is to have CodeStubAssembler be the sole assembler-like user of
the TF compiler pipeline; with RMA being a private implementation detail
and FAA being a client.)
BUG=chromium:508898
LOG=Y
Review URL: https://codereview.chromium.org/1674633002
Cr-Commit-Position: refs/heads/master@{#34852}
Add EqualStub and NotEqualStub, based on the CodeStubAssembler, and hook
them up with TurboFan and Ignition. The stubs are a full implementation
of abstract equality for ES6 plus the current SIMD.js draft, unlike the
generic version of the CompareIC, which only implements a subset with
funky runtime fallbacks.
Drive-by-fix: Introduce some common helper methods.
R=epertoso@chromium.org
BUG=chromium:592690
LOG=n
Review URL: https://codereview.chromium.org/1795793002
Cr-Commit-Position: refs/heads/master@{#34806}
Add StringLessThanStub, StringLessThanOrEqualStub, StringGreaterThanStub
and StringGreaterThanOrEqualStub, based on the CodeStubAssembler, and
hook them up with TurboFan (and Ignition). The stubs are currently
essentially comparable with the StringCompareStub, which is now
obsolete. We can later extend these stubs to cover more interesting
cases (i.e. two byte sequential string comparisons, etc.).
R=epertoso@chromium.org
Review URL: https://codereview.chromium.org/1765823002
Cr-Commit-Position: refs/heads/master@{#34485}
This adds new code stubs for abstract relational comparison,
namely LessThanStub, LessThanOrEqualStub, GreaterThanStub and
GreaterThanOrEqualStub, and hooks them up for Ignition and TurboFan.
These stubs implement the full compare operation without any
unpredictable bailouts. Currently they still go to C++ for string
comparisons, and also use the %ToPrimitive_Number runtime entry, as
we still lack a stub for the ToPrimitive operation. These issues
will be addressed separately in follow-up CLs.
Drive-by-fix: Add support for deferred code in the RawMachineAssembler
and CodeStubAssembler. A block can be marked as deferred by marking its
Label as deferred, which will then make the register allocator penalize
this block and prefer better register assignments for the other blocks.
R=epertoso@chromium.org
Review URL: https://codereview.chromium.org/1759133002
Cr-Commit-Position: refs/heads/master@{#34463}
These new stubs perform exactly the same job as the string equality case
for the CompareIC, but are platform independent and usable outside of
fullcodegen and Crankshaft. We use them in the StrictEqualStub and the
StrictNotEqualStub instead of falling back to the runtime immediately
for String comparisons, and we also use them in TurboFan to perform
String equality or inequality comparisons.
These stubs currently handle only internalized and one byte strings w/o
going to C++, but it should be easy to add support for more string cases
later, i.e. utilizing already flattened cons strings or comparing two
byte strings as well.
Review URL: https://codereview.chromium.org/1761823002
Cr-Commit-Position: refs/heads/master@{#34459}
The macro is currently used by AllocateHeapNumberStub and AllocateMutableHeapNumberStub, which are now turbofan code stubs.
It can be used to allocate objects in the new or old space, optionally with double alignment.
BUG=588692
LOG=y
Review URL: https://codereview.chromium.org/1735803003
Cr-Commit-Position: refs/heads/master@{#34424}
Initial version of a new StrictEqualStub written as TurboFan code stub,
that implements the full strict equality comparison and is usable for
both TurboFan and Ignition (and soon for the generic CompareIC case
too). The stub is not fully optimized yet, i.e. we still go to the
runtime for string comparisons, but that'll be addressed in a follow-up
CL.
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/1753173003
Cr-Commit-Position: refs/heads/master@{#34423}
Similar to fullcodegen, Ignition now also marks a for-in statement as
slow (via the TypeFeedbackVector) when we have to call %ForInFilter,
i.e. we either have no enumeration cache or the receiver map changes
during an iteration of the for-in map.
R=mstarzinger@chromium.org
BUG=v8:3650
LOG=n
Review URL: https://codereview.chromium.org/1755563002
Cr-Commit-Position: refs/heads/master@{#34391}
Operations on word size data must be word sized, and not word32.
Currently this only generates worse code, but in the future, it
might even generate wrong code, so we should better get this right
from the beginning.
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/1748953004
Cr-Commit-Position: refs/heads/master@{#34378}
Rename the existing (patching) ToBooleanStub to ToBooleanICStub to match
our naming convention, and add a new TurboFan-powered ToBooleanStub,
which just does the ToBoolean conversion without any runtime call or
code patching, so we can use it for Ignition (and TurboFan).
Drive-by-fix: Add an Oddball::to_boolean field similar to the ones we
already have for to_string and to_number, so we don't need to actually
dispatch on the concrete Oddball at all.
R=epertoso@chromium.org, rmcilroy@chromium.org, yangguo@chromium.org
Review URL: https://codereview.chromium.org/1744163002
Cr-Commit-Position: refs/heads/master@{#34361}
The ForInStep bytecode is essentially a (guaranteed) Smi increment
operation. We can do not need to go to the runtime for this operation.
R=oth@chromium.org
Review URL: https://codereview.chromium.org/1738823002
Cr-Commit-Position: refs/heads/master@{#34289}
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}
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=v8:4280
LOG=N
Review URL: https://codereview.chromium.org/1673333004
Cr-Commit-Position: refs/heads/master@{#33873}
After this change, the functionality of the CodeStubAssembler should be
sufficient to generate non-trivial stubs (e.g. the KeyedLoadIC) with control
flow, variables and probing of internal meta data structures.
Specifically this patch:
* introduces a Label class, which allows stubs to construct graphs that don't
have linear control graphs.
* introduces a Variable class. Variables can be bound to Node* values at
different points in a non-linear control flow graph. In conjunction with the
Label machinery, the CodeStubAssembler ensures that Phi nodes are inserted at
the "minimal" set of merge points.
* adds Tail calling support to other Stubs and to any arbitrary code whose
interface can be described by a CallInterfaceDescriptor.
* provides new macros for accessing FixedArray elements that are optimized for
use with Smi values.
Review URL: https://codereview.chromium.org/1649723002
Cr-Commit-Position: refs/heads/master@{#33664}
The motivation for this is that CompilationInfo really shouldn't
explicitly know anything about CodeStubs. This is evident in
the TurboFan stubs pipeline, which only needs to pass down
information about Code::Flags to the code generator and not
any of the CallInterfaceDescriptor silliness that Hydrogen has
to push around, since TF has the Linkage class that
encapsulates everything that is needed for the stub ABI. So,
instead of threading CodeStub machinery through the TF stub
pipeline, it is now removed from CompilationInfo and replaced
by only the explicit bits needed both by the Crankshaft and
TF pipelines in code generation.
Review URL: https://codereview.chromium.org/1604543002
Cr-Commit-Position: refs/heads/master@{#33410}
MachineType is now a class with two enum fields:
- MachineRepresentation
- MachineSemantic
Both enums are usable on their own, and this change switches some places from using MachineType to use just MachineRepresentation. Most notably:
- register allocator now uses just the representation.
- Phi and Select nodes only refer to representations.
Review URL: https://codereview.chromium.org/1513543003
Cr-Commit-Position: refs/heads/master@{#32738}
This moves the proper handling for the end node withing the constructed
graph into the RawMachineAssembler. This simplifies all assemblers and
makes the handling of {Start} and {End} symmetrical.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/1493963003
Cr-Commit-Position: refs/heads/master@{#32563}
* Add a sibling interface to InterpreterAssembler called
CodeStubAssembler which provides a wrapper around the
RawMachineAssembler and is intented to make it easy to build
efficient cross-platform code stubs. Much of the implementation
of CodeStubAssembler is shamelessly stolen from the
InterpreterAssembler, and the idea is to eventually merge the
two interfaces somehow, probably moving the
InterpreterAssembler interface over to use the
CodeStubAssembler. Short-term, however, the two interfaces
shall remain decoupled to increase our velocity developing the
two systems in parallel.
* Implement the StringLength stub in TurboFan with the new
CodeStubAssembler. Replace and remove the old Hydrogen-stub
version.
* Remove a whole slew of machinery to support JavaScript-style
code stub generation, since it ultimately proved unwieldy,
brittle and baroque. This cleanup includes removing the shared
code stub context, several example stubs and a tangle of build
file changes.
BUG=v8:4587
LOG=n
Review URL: https://codereview.chromium.org/1475953002
Cr-Commit-Position: refs/heads/master@{#32508}
