When vector based stores are on, we don't need to do this anymore.
BUG=
Review URL: https://codereview.chromium.org/1314433004
Cr-Commit-Position: refs/heads/master@{#30401}
A few benchmarks, e.g. Massive/SQLite, turn out to be
sensitive to compile time. Upon analysis, splinter merging
and then splinter creation (splitting) appear to be the main
contributors to such regressions. This change tackles main
sources of regression in Merging. Profiling SQLite shows,
after this change, Merging as noise (down from main C++
contributor of samples)
BUG=chromium:1318893002
LOG=n
Review URL: https://codereview.chromium.org/1318893002
Cr-Commit-Position: refs/heads/master@{#30400}
While the intention is to eventually do away with
FLAG_turbo_loop_stackcheck and FLAG_turbo_preprocess_range,
they are useful for the interim we are still testing and
benchmarking the feature.
Review URL: https://codereview.chromium.org/1314163003
Cr-Commit-Position: refs/heads/master@{#30399}
Port cbbaf9ea6a
Note that the above commit breaks embedded constant pools and will need to
be revised in a future CL.
Original commit message:
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.
R=danno@chromium.org, titzer@chromium.org, jyan@ca.ibm.com, dstence@us.ibm.com, joransiu@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/1321553002
Cr-Commit-Position: refs/heads/master@{#30396}
TC39 agreed to disallow "use strict" directives in function body when
non-simple parameter lists are used.
This is a continuation of caitp's CL https://codereview.chromium.org/1281163002/
with some refactorings removed for now.
Still TODO: there is a lot of duplication between the is_simple field of
FormalParametersBase and the NonSimpleParameter property ExpressionClassifier
keeps track of. It should be possible to remove the former with a minor
refactoring of arrow function parsing. This will be attempted in a follow-up CL.
BUG=
LOG=N
Review URL: https://codereview.chromium.org/1300103005
Cr-Commit-Position: refs/heads/master@{#30388}
When encountering a "=" token in ParseAssignmentExpression, the default
parameter case is not locally distinguishable from the destructuring case.
BUG=
Review URL: https://codereview.chromium.org/1317843002
Cr-Commit-Position: refs/heads/master@{#30386}
%_IsObject(foo) is equivalent to typeof foo === 'object' and has
exactly the same optimizations, so there's zero need for %_IsObject
in our code base.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/1313903003
Cr-Commit-Position: refs/heads/master@{#30380}
For vector-based keyed store ics, we need to know the current
KeyedAccessStore mode on ic MISS, and to produce optimized code.
We can't store this mode, which can change on any MISS in the IC
without patching. Therefore, this CL makes sure that the information is
redundantly available in the handlers embedded in the IC. This way,
when --vector-stores is turned on, we'll be able to extract that
information from the vector which maintains a list of these handlers.
BUG=
Review URL: https://codereview.chromium.org/1312693004
Cr-Commit-Position: refs/heads/master@{#30378}
port 5d875a57fa (r30342).
original commit message:
The previous hack with HInstanceOfKnownGlobal was not only slower,
but also very brittle and required a lot of weird hacks to support it. And
what's even more important it wasn't even correct (because a map check
on the lhs is never enough for instanceof).
The new implementation provides a sane runtime implementation
for InstanceOf plus a fast case in the InstanceOfStub, combined with
a proper specialization in the case of a known global in CrankShaft,
which does only the prototype chain walk (coupled with a code
dependency on the known global).
As a drive-by-fix: Also fix the incorrect Object.prototype.isPrototypeOf
implementation.
R=weiliang.lin@intel.com
BUG=
Review URL: https://codereview.chromium.org/1318663003
Cr-Commit-Position: refs/heads/master@{#30376}
This fixes broken dynamic hole-checks for the temporal dead zone of
non-initializing assignments to {let} and {const} declared variables.
Also note that this exemplifies a case where the dynamic check for such
assignments to {let} declared variables can no longer be elided as the
comment suggested.
R=rossberg@chromium.org
TEST=mjsunit/regress/regress-4388
BUG=v8:4388
LOG=n
Review URL: https://codereview.chromium.org/1318693002
Cr-Commit-Position: refs/heads/master@{#30375}
A TopLevelLiveRange is the live range of a virtual register. Through
register allocation, it may end up being split in a succession of child
live ranges, where data flow is handled through moves from
predecessor to successor child.
Today, the concepts of "top level" and "child" live ranges are conflated
under the LiveRange class. However, a good few APIs pertain solely
to TopLevelLiveRanges. This was communicated through comments or
DCHECKs - but this makes for poor code comprehensibility and maintainability.
For example, the worklist of the register allocator (live_ranges()) needs
to only contain TopLevelLiveRanges; spill range concerns are associated
only with the top range; phi-ness; certain phases in the allocation pipeline;
APIs on LiveRange used for initial construction - before splitting;
splintering - these are all responsibilities associated to TopLevelLiveRanges,
and not child live ranges.
This change separates the concepts.
An effect of this change is that child live range allocation need not involve
RegisterAllocationData. That's "a good thing" (lower coupling), but it has
the side-effect of not having a good way to construct unique identifiers for
child live ranges, relative to a given InstructionSequence.
LiveRange Id are used primarily for tracing/output-ing, and debugging.
I propose a 2-component identifier: a virtual register (vreg) number,
uniquely identifying TopLevelLiveRanges; and a relative identifier, which
uniquely identifies children of a given TopLevelLiveRange. "0" is reserved
for the TopLevel range. The relative identifier does not necessarily
indicate order in the child chain, which is no worse than the current state
of affairs.
I believe this change should make it easier to understand a trace output
(because the virtual register number is readily available). I plan to formalize
with a small structure the notion of live range id, and consolidate tracing
around that, as part of a separate CL. (there are seemingly disparate ways
to trace - printf or stream-based APIs - so this seems like an opportune
change to consolidate that)
Review URL: https://codereview.chromium.org/1311983002
Cr-Commit-Position: refs/heads/master@{#30370}
For X87 platform, it has the same general register as ia32 and it will spill the
context to the stack too.
port bfdc22d7fc (r29325).
original commit message:
Passes the current context to bytecode interpreter handlers. This is held in the
context register on all architectures except for ia32 where there are too few
registers and it is instead spilled to the stack.
Also changes Load/StoreRegister to use kMachAnyTagged representation since they
should only ever hold tagged values.
BUG=
Review URL: https://codereview.chromium.org/1316583003
Cr-Commit-Position: refs/heads/master@{#30368}
The lack of marking this dependency led to a ClusterFuzz crash when
sloppy-function was on but not sloppy. This case does not make sense.
R=adamk
LOG=N
BUG=chromium:520891
Review URL: https://codereview.chromium.org/1316773004
Cr-Commit-Position: refs/heads/master@{#30364}
Make ObjectLiteral::is_simple() false for literals containing computed
property names, which causes IsCompileTimeValue() to return false and
thus force code to be generated for setting up such properties. This
mirrors the handling of '__proto__' in literals.
BUG=v8:4387
LOG=y
Review URL: https://codereview.chromium.org/1307943007
Cr-Commit-Position: refs/heads/master@{#30362}
This change encompasses what is necessary to enable stack checks in loops without suffering large regressions.
Primarily, it consists of a new mechanism for dealing with deferred blocks by "splintering", rather than splitting, inside deferred blocks.
My initial change was splitting along deferred block boundaries, but the regression introduced by stackchecks wasn't resolved conclusively. After investigation, it appears that just splitting ranges along cold block boundaries leads to a greater opportunity for moves on the hot path, hence the suboptimal outcome.
The alternative "splinters" ranges rather than splitting them. While splitting creates 2 ranges and links them (parent-child), in contrast, splintering creates a new independent range with no parent-child relation to the original. The original range appears as if it has a liveness hole in the place of the splintered one. All thus obtained ranges are then register allocated with no change to the register allocator.
The splinters (cold blocks) do not conflict with the hot path ranges, by construction. The hot path ones have less pressure to split, because we remove a source of conflicts. After allocation, we merge the splinters back to their original ranges and continue the pipeline. We leverage the previous changes made for deferred blocks (determining where to spill, for example).
Review URL: https://codereview.chromium.org/1305393003
Cr-Commit-Position: refs/heads/master@{#30357}
