This CL attempts to simplify the Assembler's dependency on the
isolate, in particular on a global "serializer_enabled" mode contained
therein. The "serializer_enabled" condition enabled and disabled
a number of things in both the assemblers and macro assemblers. To
make these dependencies explicit, the Assembler::IsolateData is refactored
to be a proper Assembler::Options struct that controls specific assembler
behaviors, with default settings easily computable from the isolate.
This also helps make the contract for compiling WASM code more explicit
(since WASM code needs to have reloc info recorded for external references)
we can explicitly enable this recording without trying to "trick" the
assembler using "serializer_enabled".
R=jgruber@chromium.orgCC=mstarzinger@chromium.org, herhut@chromium.org
Change-Id: I7a8ba49df7b75b292d73ec2aa6e507c27a3d99c8
Reviewed-on: https://chromium-review.googlesource.com/1105982
Commit-Queue: Ben Titzer <titzer@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53890}
When encountering a LoadStackPointer input to a comparison, generate a register
LocationOperand that points to the stack pointer. This can avoid unnecessary
spilling of the stack pointer.
Since sp is a special register for arm64, we need to add a mechanism to print
its name in RegisterConfiguration.
This is a port of https://chromium-review.googlesource.com/1055568 that made
the same change for arm.
It also ports the tests added in
https://chromium-review.googlesource.com/1099068 to arm and arm64.
Bug: v8:7844
Change-Id: I5adc672ff877b9888ef755e8e60e4eabbc61061b
Reviewed-on: https://chromium-review.googlesource.com/1107810
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Cr-Commit-Position: refs/heads/master@{#53889}
There is no point calling AllocationSpaceName manually every time. We
could get rid of AllocationSpaceName entirely, except that the
serializer calls it directly on ints that are casted to AllocationSpace
objects first.
Change-Id: I26b2181347dba960c8523d5dcecf8128208c59a2
Reviewed-on: https://chromium-review.googlesource.com/1107816
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53888}
Some cctests force fresh creation of heap constants, even though the
cctest binary itself is an embedded snapshot build (i.e.: a snapshot
blob exists, and a binary-embedded blob exists). This breaks a few
assumptions, for example that off-heap builtins have a single,
canonical off-heap code range.
Unfortunately this isn't that easy to fix. I see a few alternatives:
1. In builtins setup, if an embedded blob exists, regenerate the
builtins for their metadata (things like the safepoint table offset),
and then replace them by off-heap trampolines.
2. As above, but deserialize the trampolines from the snapshot blob.
3. As above, but pack required metadata into the embedded blob and
create trampolines from there.
4. Act as if the embedded blob does not exist.
Alternative 1 does not work because the generated code can be slightly
different at at runtime vs. mksnapshot-time. Alternative 2 is out
because we do not have access to the snapshot blob in TestIsolate
setup. Alternative 3 is probably the preferred option but would be a
more involved change.
This CL takes path 4. It's not an optimal solution, but it can be
replace by alternative 3 later.
TBR=ulan@chromium.org
Bug: v8:7718, v8:7751
Change-Id: I36c024cb0179615011c886ed3978bc95f0d197ac
Reviewed-on: https://chromium-review.googlesource.com/1098924
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53886}
The stack check builtin really has stub linkage, not wasm linkage. For
the stack check, it technically does not make a difference currently
since the stack check does not receive any arguments.
R=mstarzinger@chromium.org
Change-Id: Ia3e6ea192250dc3799c7a5166cf616f883c1f63f
Reviewed-on: https://chromium-review.googlesource.com/1107709
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53883}
Unconditionally setting the JSFunction code to that of the SFI
may skip initializing the feedback vector.
R=leszeks@chromium.org
Bug: v8:7857
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I65d4bf32493be4cade2eaf3d665d44f93e80f809
Reviewed-on: https://chromium-review.googlesource.com/1107618
Commit-Queue: Yang Guo <yangguo@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53881}
This test was not adapted for the jump table yet.
R=mstarzinger@chromium.org
Bug: v8:7867
Change-Id: Ifd03899b0757972525dcc54f04949635d03fa493
Reviewed-on: https://chromium-review.googlesource.com/1107624
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53878}
If the isolate is available via another object (including JSReceivers
and Contexts), grab it through that rather than HeapObject::GetIsolate.
As a drive-by, fix up places where this looked like it would work but
didn't because of null or uninitialized objects.
Bug: v8:7786
Change-Id: I3454840073d1455b3738e0ac34ccaa3c9a220896
Reviewed-on: https://chromium-review.googlesource.com/1104170
Reviewed-by: Dan Elphick <delphick@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53876}
Replace all implicit Handle<T>(T*) constructors in src/bootstrapper.cc
with explicit ones that use a local variable or parameter.
(All done using Leszek's awesome new tool!)
Bug: v8:7786
Change-Id: Iec01fda502dd121bc1815d5dab73bd09bb3e5aa6
Reviewed-on: https://chromium-review.googlesource.com/1107697
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53873}
This is a reland of 380dba0a5c
Original change's description:
> [runtime][gc] Drop AllocationSite::weak_next field for sub-literals
>
> Use AllocationSite without Weaknext field for all the allocations in nested
> literal except for Root. The nested field is sufficient to link all the
> allocations in a nested literal. Only the Root is added to heap weak_alloc_list
> for GC to traverse
>
> Change-Id: I946e63292c6d168197cd2a087f697c73cc431272
> Reviewed-on: https://chromium-review.googlesource.com/1101323
> Commit-Queue: Chandan Reddy <chandanreddy@google.com>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#53813}
Change-Id: Ic451fb3600b87f8f69748b1659ff89a80c9322fa
Reviewed-on: https://chromium-review.googlesource.com/1106257
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Chandan Reddy <chandanreddy@google.com>
Cr-Commit-Position: refs/heads/master@{#53872}
There's no need to assume anything when we don't reduce the node.
R=bmeurer@chromium.org
Change-Id: Ic39b6e634c662824a72affd91318ca024f3e1055
Reviewed-on: https://chromium-review.googlesource.com/1105832
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53870}
This CL adds a constexpr int31 == operator that returns
a constexpr bool value.
It also fixes a bug that this introduces because branch generation
for ternary operators incorrectly assumed that the condition value
was bool, whereas it can also be constexpr bool now.
Change-Id: Ie7d3497c9d1112c121b4a29561fe065b5bf01fb6
Reviewed-on: https://chromium-review.googlesource.com/1106343
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Théotime Grohens <theotime@google.com>
Cr-Commit-Position: refs/heads/master@{#53867}
This is in order to check if we can get rid of the cache.
Bug: v8:7790
Change-Id: Id6625ffc440665b090f999564b006506f947ae45
Reviewed-on: https://chromium-review.googlesource.com/1107620
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53865}
We can instead pass the deopt id in a register, where before we were passing the
deopt entry address. This removes the need for the deopt tables altogether,
saving 192kB.
Change-Id: I479d4de1a0245de328720b6b03a1955c8c63f696
Reviewed-on: https://chromium-review.googlesource.com/1076472
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Cr-Commit-Position: refs/heads/master@{#53863}
Also fix an oversight in my previous CL.
R=jarin@chromium.org
Bug: v8:7790
Change-Id: I61c783392b7b7b38ea28dc44dc1e932d15b55bc6
Reviewed-on: https://chromium-review.googlesource.com/1106170
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53862}
GCC 4.9 used on some Node.js CI machines complains when the control
reaches the end of a non-void function and no return is encountered.
R=bmeurer@google.com, ofrobots@google.com, yangguo@google.com
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I5af0192cb187eccbf34dbb60ff3ac2e4774af803
Reviewed-on: https://chromium-review.googlesource.com/1105619
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53861}
The reference to the {NativeModule} (stored in a {Managed}) should live
on the {WasmModuleObject}, not on the individual {WasmCompiledModule}
objects.
R=titzer@chromium.org
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I21dfa088c9643d36d9fd5052a145e7e2af5e47f9
Reviewed-on: https://chromium-review.googlesource.com/1106380
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53860}
The {ModuleEnv} already contains a pointer to the {WasmModule}, no need
to pass it explicitly.
R=titzer@chromium.org
Change-Id: Icf0e8ea8b25c33dd5bcaeab2a4a746376e73813d
Reviewed-on: https://chromium-review.googlesource.com/1105828
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53858}
Our tests currently don't set {WasmModule::num_declared_functions}
correctly. This CL fixes that.
This enables the use of {WasmModule::num_declared_functions} instead of
{NativeModule::num_functions_ - NativeModule::num_imported_functions_}.
Drive-by: Fix {std::vector} reservation to reserve enough space for all
functions during decoding.
R=titzer@chromium.org
Change-Id: I6d7783aed1c0de3275fc72787dec17c38ff8c73b
Reviewed-on: https://chromium-review.googlesource.com/1106166
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53857}
Port 733b7c8258
Original Commit Message:
This introduces the concept of a jump table for WebAssembly, which is
used for every direct and indirect call to any WebAssembly function.
For lazy compilation, it will initially contain code to call the
WasmCompileLazy builtin, where it passes the function index to be
called.
For non-lazy-compilation, it will contain a jump to the actual code.
The jump table allows to easily redirect functions for lazy
compilation, tier-up, debugging and (in the future) code aging. After
this CL, we will not need to patch existing code any more for any of
these operations.
R=clemensh@chromium.org, joransiu@ca.ibm.com, michael_dawson@ca.ibm.com
BUG= v8:7758
Change-Id: I1a370910ffa56dbdd609be6922067842a6adf1df
Reviewed-on: https://chromium-review.googlesource.com/1105060
Commit-Queue: Junliang Yan <jyan@ca.ibm.com>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53852}
This removes special casing in the handling of {RelocInfo::CODE_TARGET}
for cases where such references appeared in {WasmCode}. Such references
can only appear in real {Code} objects and also only target real {Code}
objects by now.
R=clemensh@chromium.org
Change-Id: I87c55c60cd6344bb67fa8c8d04b66fb523d07ba7
Reviewed-on: https://chromium-review.googlesource.com/1104680
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53848}
Direct jumps only work for offset up to 64 MB on arm. For longer jumps,
use indirect branches (load target from constant pool into the pc
register).
R=mstarzinger@chromium.orgCC=pierre.langlois@arm.com
Bug: v8:7758
Change-Id: I1cf66b7d1bfb62cfcd6b1619c02816909a1f651e
Reviewed-on: https://chromium-review.googlesource.com/1105996
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53847}
This makes sure we use WebAssembly runtime stubs when performing value
coercions in wrapper code that is compiled into {WasmCode}. It hence
removes the last {RelocInfo::CODE_TARGET} references in {WasmCode}.
R=clemensh@chromium.org
BUG=v8:7424
Change-Id: Ic96e541abe98f90796a6506b09ff99557743b6f7
Reviewed-on: https://chromium-review.googlesource.com/1104462
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53846}
This encapuslates some of the heap accesses done by
JSNativeContextSpecialization::ReduceJSLoadGlobal and
JSNativeContextSpecialization::ReduceJSStoreGlobal.
Bug: v8:7790
Change-Id: Ib6c63903809927d6094af22519285cb9d0bbff7a
Reviewed-on: https://chromium-review.googlesource.com/1106141
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53845}
After the removal of the --skip-compiling-wasm-funcs flag, this helper
does not help that much any more, and contains unneeded logic. This CL
just inlines the computation in the caller.
R=titzer@chromium.org
Bug: v8:7754
Change-Id: I99cf006ffe9c33d79bea2ce186354c38bc63f6af
Reviewed-on: https://chromium-review.googlesource.com/1105767
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53844}
Like Heap::InFromSpace, it doesn't use any members, so should be made
static. Also clean up call sites to not call via a heap pointer.
Change-Id: If55484ddac51351d789c73093f1f7ebf1c568bea
Reviewed-on: https://chromium-review.googlesource.com/1088618
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53842}
