This reverts commit 1280954d3a.
Reason for revert: Speculative, GC stress bots started taking much longer after this change.
Original change's description:
> [wasm] Throttle the amount of unfinished work to avoid OOM
>
> It is possible that the foreground task is unable to clear the
> scheduled unfinished work, eventually leading to an OOM.
>
> We use either code_range on 64 bit, or the capacity of the code space,
> as a heuristic for how much memory to use for compilation.
>
> Bug: v8:6492, chromium:732010
> Change-Id: I1e4c0825351a42fa0b8369ccc41800ac3445563d
> Reviewed-on: https://chromium-review.googlesource.com/535017
> Commit-Queue: Brad Nelson <bradnelson@chromium.org>
> Reviewed-by: Brad Nelson <bradnelson@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#46017}
TBR=bradnelson@chromium.org,mtrofin@chromium.org,ahaas@chromium.org
Change-Id: I8883cee7f77667530bc50f91bfb468c485e6f7f2
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:6492, chromium:732010
Reviewed-on: https://chromium-review.googlesource.com/540270
Reviewed-by: Bill Budge <bbudge@chromium.org>
Commit-Queue: Bill Budge <bbudge@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46020}
Adds a 'performance' test which tracks the number of expressions
which can be nested before the compiler runs out of stack space.
This isn't really a performance test, but is created as a js-perf-test
to enable regression tracking in the dashboards.
Change-Id: Iee0c00df53b38b083e2dde09676ac9b13e439461
Reviewed-on: https://chromium-review.googlesource.com/539419
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46019}
It is possible that the foreground task is unable to clear the
scheduled unfinished work, eventually leading to an OOM.
We use either code_range on 64 bit, or the capacity of the code space,
as a heuristic for how much memory to use for compilation.
Bug: v8:6492, chromium:732010
Change-Id: I1e4c0825351a42fa0b8369ccc41800ac3445563d
Reviewed-on: https://chromium-review.googlesource.com/535017
Commit-Queue: Brad Nelson <bradnelson@chromium.org>
Reviewed-by: Brad Nelson <bradnelson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46017}
- Use correct prefixes for SIMD/Atomics ops
- S128 LoadMem/StoreMem should not use 0xc0/0xc1 opcodes, these are now
being used for sign extension
- S128 LoadMem/StoreMem should use prefixed opcodes
BUG=v8:6020
Review-Url: https://codereview.chromium.org/2943773002
Cr-Commit-Position: refs/heads/master@{#46016}
This patch updates the error positition and the error msg.
Previously,
→ ./out.gn/x64.release/d8 test.js
test.js:1: TypeError: undefined is not a function
var [a] = {};
^
TypeError: undefined is not a function
at test.js:1:1
With this patch,
→ ./out.gn/x64.release/d8 test.js
test.js:1: TypeError: [Symbol.iterator] is not a function
var [a] = {};
^
TypeError: [Symbol.iterator] is not a function
at test.js:1:11
Bug: v8:5532
Change-Id: Ib066e8ec8a53fdf06cce491bde4b1d0c6d564cbc
Reviewed-on: https://chromium-review.googlesource.com/539024
Reviewed-by: Adam Klein <adamk@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46015}
Merge OLD_TO_OLD and OLD_TO_NEW per page. This enables removing atomic
operations for the slot updates, effectively removing the need for
fences.
Bug: chromium:651354
Change-Id: I9e318bef06c403b135d638cf94fda9569dcf0e36
Reviewed-on: https://chromium-review.googlesource.com/539338
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46014}
For functions without any calls, there is no value in executing a stack
check. The current frame is materialized at that point anyway.
Note that for loops, we still emit additional stack checks in the loop
header.
For unity, the reduction in code size is moderate (0.53%), as only 4000
of the 34000 functions are leaf functions (no calls). However, we also
save some compile time and gain performance, so this is still worth
doing it.
Drive-by: Fix the effect chain generated in {StackCheck()}.
R=mstarzinger@chromium.org, ahaas@chromium.org
Change-Id: Ia6ec58d0ea46de02634c923cdf8e6e08d8902c59
Reviewed-on: https://chromium-review.googlesource.com/533333
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46013}
We only need to use this for certain Intrinsics defined in the spec.
This CL removes unnecessary uses.
Bug: v8:6474
Change-Id: I13a9f0c57d877dd65a883a38f9683d55623030d3
Reviewed-on: https://chromium-review.googlesource.com/529224
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46012}
Track execution counts of the continuations of block structures (e.g.
IfStatements) to capture cases in which execution does not continue after a
block. For example:
for (;;) {
return;
}
// Never reached, tracked by continuation counter.
A continuation counter only has a start position; it's range is implicitly
until the next sibling range or the end of the parent range.
Bug: v8:6000
Change-Id: I8e8f1f5b140b64c86754b916e626eb50f0707d70
Reviewed-on: https://chromium-review.googlesource.com/530846
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46006}
If we pass in thin-string into a keyed load, the underlying internalized string is used to find the handler. However, the thin string itself was used to interpret the handler. Since the thin string itself isn't unique, this caused existing properties on the prototype chain to not be found in case of dictionary-mode prototypes.
Bug: chromium:731193
Change-Id: Ic98d3789ecf9175e17d9c898ab13231aad59efcc
Reviewed-on: https://chromium-review.googlesource.com/539596
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46005}
It might happen that we deprecate the map of previous sub-literals if we create
literals with the same map several times. This is usually the case for
configuration arrays.
Bug: chromium:734051
Change-Id: I82284e5aae632286135b2092816d776d229c65af
Reviewed-on: https://chromium-review.googlesource.com/538665
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46003}
Here we optimize Dsubu by instead of loading imm and subtracting, we
load -imm and perform addition when loading -imm takes less instructions
than loading imm. Similarily li is optimized by loading -imm and
performing addition or loading ~imm and inverting bits using nor when
one of these loads takes two instructions less than loading imm, saving
at least one instruction. Tests are adjusted to cover these
optimizations.
BUG=
TEST=cctest/test-assembler-mips/li_macro
cctest/test-assembler-mips/Dsubu
Review-Url: https://codereview.chromium.org/2909913002
Cr-Commit-Position: refs/heads/master@{#46001}
Mechanical change to remove argument adaption (should be a tad faster
this way). Especially next is called without arguments in the common
case.
Bug: v8:6354, v8:6369
Change-Id: I4180caabfc4c1bbf1a10a881dcbcd41e03614b27
Reviewed-on: https://chromium-review.googlesource.com/535453
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Caitlin Potter <caitp@igalia.com>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46000}
Large allocations would fail due to the flag not being set.
Bug: chromium:732836
Change-Id: I31686e382386a2d08582c86b29dc8f89841040d1
Reviewed-on: https://chromium-review.googlesource.com/535563
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45999}
Minor differences in how we dispatch on the regexp type
(IRREGEXP,ATOM,NOT_COMPILED) make significant differences in benchmark
performance. A simple switch turns out to be the best alternative.
BUG=chromium:734035, v8:6462
Change-Id: I09c613658e828b9fd1e3082624ef692b8b4a0c5f
Reviewed-on: https://chromium-review.googlesource.com/539295
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45998}
For interpreted functions, use the optimized code slot in the feedback
vector to store an optimization marker (optimize/in optimization queue)
rather than changing the JSFunction's code object. Then, adapt the
self-healing mechanism to also dispatch based on this optimization
marker. Similarly, replace SFI marking with optimization marker checks
in CompileLazy.
This allows JSFunctions to share optimization information (replacing
shared function marking) without leaking this information across native
contexts. Non I+TF functions (asm.js or --no-turbo) use a
CheckOptimizationMarker shim which generalises the old
CompileOptimized/InOptimizationQueue builtins and also checks the same
optimization marker as CompileLazy and InterpreterEntryTrampoline.
This is a reland of https://chromium-review.googlesource.com/c/509716
Change-Id: I02b790544596562373da4c9c9f6afde5fb3bcffe
Reviewed-on: https://chromium-review.googlesource.com/535460
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45997}
R=neis@chromium.org
Change-Id: I23298e2c0adcfdc4e6e963e98cde641bef9cdb5b
Reviewed-on: https://chromium-review.googlesource.com/539296
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45996}
Unify, simplify logic, reduce UTF8 specific handling.
Intend of this is also to have stream views.
Stream views can be used concurrently by multiple threads, but
only one thread may fetch new data from the underlying source.
This together with unified stream view creation is intended to be
used for parse tasks.
BUG=v8:6093
Change-Id: I83c6f1e6ad280c28da690da41c466dfcbb7915e6
Reviewed-on: https://chromium-review.googlesource.com/535474
Reviewed-by: Daniel Vogelheim <vogelheim@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Commit-Queue: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45994}
Toon suggested this as a mitigation to the problem of prototype fast mode switching invalidating prototype chain validity cell, and thus sending keyed store ICs to megamorphic state.
BUG=chromium:723479
Review-Url: https://codereview.chromium.org/2943313002
Cr-Commit-Position: refs/heads/master@{#45993}
This CL also reduces the minimum semi-space size to 512K.
BUG=chromium:716032
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_rel_ng
Review-Url: https://codereview.chromium.org/2942543002
Cr-Commit-Position: refs/heads/master@{#45992}
This removes both {BailoutId} as well as {TypeFeedbackId} numbers from
almost all AST nodes. The only exception are {IterationStatement} nodes
which still require an ID for on-stack replacement support.
R=verwaest@chromium.org
BUG=v8:6409
Change-Id: I5f7b7673ae5797b9cbc9741144d304f0d31d4446
Reviewed-on: https://chromium-review.googlesource.com/538792
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45991}
... in order to avoid creating an OsrHelper during code assembly,
because its constructor accesses the heap.
Bug: v8:6048
Change-Id: I3bf592a5a0f91752a9f5ec35982f962445512bb7
Reviewed-on: https://chromium-review.googlesource.com/530370
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45990}
We need to constant-fold JSHasInPrototypeChain nodes early during
inlining, otherwise we already miss a couple of optimization
opportunities if we wait until after typing. This moves the
constant-folding part of the JSHasInPrototypeChain lowering back to
JSNativeContextSpecialization, where it was before the changes in
https://codereview.chromium.org/2934893002 (part of
JSOrdinaryHasInstance lowering back then).
BUG=v8:5269,v8:5989,v8:6483,chromium:733158
R=jgruber@chromium.org
Review-Url: https://codereview.chromium.org/2943293002
Cr-Commit-Position: refs/heads/master@{#45989}
The heap verifier does certain invariant checks on JSBoundFunction
objects, i.e. it assumes that the bound_target_function is a proper
JSReceiver. The Deoptimizer cannot maintain this invariant, because it
first allocates the JSBoundFunction in an invalid state and only
afterwards fix up the state. But the GC (and thus the heap verifier)
can observe this invalid state why materializing field values, so
we need to relax the verification slightly.
BUG=chromium:729573,chromium:732176
R=mstarzinger@chromium.org
Review-Url: https://codereview.chromium.org/2933283002
Cr-Commit-Position: refs/heads/master@{#45988}
When iterating over stack frames in the cpu profiler, don't perform any
object casts that have heap-testing DCHECKs. Instead, access values on
the frame by offsets directly, and only check their tags for validity.
Cq-Include-Trybots: master.tryserver.chromium.linux:linux_chromium_rel_ng
Change-Id: Ia54b18f8ab947c1827f17483806104f0d1d34136
Reviewed-on: https://chromium-review.googlesource.com/536973
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45985}
This class contained a by-now unnecessary optimization of FindEntry. Since we always deal with internalized names by now anyway, there's no need to micro-optimize locally (it's a nop).
Bug:
Change-Id: I5a0046bcd23e2cb77c5902e850bac6211bd5518f
Reviewed-on: https://chromium-review.googlesource.com/538581
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45983}
The Smi versions of arithmetic bytecodes (AddSmi, SubSmi, MulSmi,
DivSmi, ModSmi) have a fast path for Smi case and call to a builtin
on the slow path. However, this builtin is only used by these bytecode
handlers. This cl removes the builtins and inlines them into
bytecode handlers. This will also save few checks in the slow-path.
Subtract, multiply, divide and modulus also share the same checks to
collect type feedback on several cases. This cl also refactors them
to share the same code.
Also removed a couple of TODOs that are no longer relevant.
Bug: v8:4280, v8:6474
Change-Id: Id23bd61c2074564a1beacb0632165f52370ff226
Reviewed-on: https://chromium-review.googlesource.com/530845
Commit-Queue: Mythri Alle <mythria@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45982}
With the introduction of the fast-cloning double fields in the CSA stub for
literals we forgot to check for deprecated maps. As a result every subsequent
IC-miss would have to migrate the objects from such boilerplates.
This CL makes sure we don't use the deprecated map when copying boilerplates,
thus restoring the original behavior.
Bug: v8:6211 chromium:728682
Change-Id: If9ea1e0c5c6fb4236cb7a82ea33306a600925ac3
Reviewed-on: https://chromium-review.googlesource.com/538677
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45981}
Remove dead code on the way.
Bug: v8:6474
Change-Id: I7edb4277bc53ee92edf9523b943492782ec6efac
Reviewed-on: https://chromium-review.googlesource.com/538652
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45976}
Storing the boilerplate on the first run leads to memory ovehead for code
that is run only once. Hence we directly return the creating literal on the
first run and only start creating copies from the second run on.
Bug: v8:6211
Change-Id: I69b96d124a5b594b991fdbcc76dbf935d973ffad
Reviewed-on: https://chromium-review.googlesource.com/530688
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45975}
Profiler ticks are reset when the type feedback changes for Load / Store ICs.
This cl extends this to other operations as well. This allows us to tier up
functions when the feedback vectors are stable. This is the first step for
a set of follow up cls that will change the heuristics used in
runtime-profiler.
Bug:
Change-Id: I875209712c6161e425a03475c14890a49155c0e1
Reviewed-on: https://chromium-review.googlesource.com/529165
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Mythri Alle <mythria@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45974}
This is in preparation for lowering monomorphic loads during graph building.
This essentially moves the parts that will be shared to a separate class/file
(proparty-access-builder.(cc|h)).
I should say that we will not want to do accessor inlining during graph
building because that would require us to create frame states
(which is the thing we would like to avoid doing).
Review-Url: https://codereview.chromium.org/2936673005
Cr-Commit-Position: refs/heads/master@{#45973}
This removes the heuristic from {JSStackFrame::IsConstructor} that tried
to infer whether a frame was called as a constructor or not from the
receiver value. We are now carrying along the appropriate bit derived
from the frame type instead.
R=jgruber@chromium.org
TEST=message/regress/regress-5727
BUG=v8:5727
Change-Id: I0e2f1d0f95485c84c4ebcd3cbfe0123c6afd2e01
Reviewed-on: https://chromium-review.googlesource.com/500313
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#45972}
