Rather than patching code, the deoptimizer now replaces the
return address in the frames with respective trampolines.
This change required to change the way we search for Safepoint
entries and for Exception Handlers.
It's working in architectures: x64, ia32, arm, arm64 and mips.
Bug: V8:6563
Change-Id: I3cbd4d192c3513f307b3a6a2ac99e60d03c753d3
Reviewed-on: https://chromium-review.googlesource.com/586707
Commit-Queue: Juliana Patricia Vicente Franco <jupvfranco@google.com>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46967}
LayoutDescriptor is currently the only case where the Scavenger needs to
potentially follow an updated slot to iterate an object. This scenario
requires at least Acq/Rel semantics.
In order to use relaxed store/load for the slots we need to allocate it
pretenured.
Bug: chromium:738865
Change-Id: I353fa6f252b436918b60fff54ece5ce5c7783072
Reviewed-on: https://chromium-review.googlesource.com/590429
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46965}
Add UseScratchRegisterScope for MIPS and use it instead of using at register directly.
Original commit message:
`Introduce a stripped down version of UseScratchRegisterScope for ARM and use it
inside the assembler and macro-assembler. At the exception of the Call
instructions, we now use this scope instead of using the ip register
directly. This is inspired from how the ARM64 backend works.
In general, the benefit of doing this is we can catch cases where ip is being
used both by the caller and by the assembler. But more specifically, TurboFan
reserves r9 as an extra scratch register because ip can already be used by the
assembler. With this utility, we can isolate the cases in the code generator
which need an extra register and potentially fix them, allowing us to give r9
back to the register allocator.
This patch uncovered places in the assembler where we were using ip
unconditionally when we could have re-used the destination register instead.`
Bug:
Change-Id: I1a35c1661579882801605337abfc95f75b47f052
Reviewed-on: https://chromium-review.googlesource.com/574923
Commit-Queue: Ivica Bogosavljevic <ivica.bogosavljevic@imgtec.com>
Reviewed-by: Ivica Bogosavljevic <ivica.bogosavljevic@imgtec.com>
Cr-Commit-Position: refs/heads/master@{#46963}
Rather than storing the isolate in compiler dispatcher jobs, which gets
weird for jobs that are entirely off-thread, instead pass the isolate
in when stepping on the main thread.
This makes it clearer which steps must be executed on the main thread,
as they require the caller to explicitly give them access to the
isolate.
Bug: v8:6537
Change-Id: I02fff7c77fdcdbfb099a38235f94d8c1040699ac
Reviewed-on: https://chromium-review.googlesource.com/589437
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46962}
With TurboFan, there should no longer be any deopt loops (aside from
bugs). So, the "too many deopts" bailout is no longer needed, at least
in its current form.
This fixes an issue where deopt counts are leaked between native
contexts, resulting in optimization being disabled unnecessarily.
Bug: v8:6402
Change-Id: Ia06374ae6b5c2d473bcdd8eef1284bf02766c2fb
Reviewed-on: https://chromium-review.googlesource.com/588894
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46961}
This includes the catch and finally keywords in the respective range.
For instance:
// Catch range previously: |<--------->|
try { /* ... */ } catch (e) { /* ... */ }
// Now: |<------------------->|
Bug: v8:6000
Change-Id: I1bd9f7fce8bb7de945da83ab512833841b9d956a
Reviewed-on: https://chromium-review.googlesource.com/586598
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46959}
Port 37680d6563
Original Commit Message:
Instead of having feedback vector as a subtype of FixedArray with
reserved slots, make it a first-class variable-sized object with a
fixed-size header. This allows us to compress counters to ints in the
header, rather than forcing them to be Smis.
R=leszeks@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=
LOG=N
Change-Id: Ia835942de292c4e4b802e34672f1e8bf8a2491c7
Reviewed-on: https://chromium-review.googlesource.com/590168
Reviewed-by: Junliang Yan <jyan@ca.ibm.com>
Commit-Queue: Jaideep Bajwa <bjaideep@ca.ibm.com>
Cr-Commit-Position: refs/heads/master@{#46954}
This reverts commit 3ca6408511.
Reason: it does not seem to improve anything (including microbenchmarks
that only do "if (m.has(x)) s += m.get(x);" in a tight loop).
Bug: v8:6410
Change-Id: I025bf885f313ac5e54ca450ae9cff5b4a15b04fd
Reviewed-on: https://chromium-review.googlesource.com/574020
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46952}
Everything inside a class lives inside the class scope, so
reparenting the class scope is the only operation that
should be done to ClassLiterals during reparenting.
Bug: chromium:740591
Change-Id: Ia5b96b44ff1ca6cfa274effb5a04651809bab9bd
Reviewed-on: https://chromium-review.googlesource.com/588054
Commit-Queue: Adam Klein <adamk@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46951}
Port 79bcb45447
Original Commit Message:
This is a reland of a72b2f88a8
Original change's description:
> [arm] Restrict grouping pushes before a TailCall to registers only
>
> We optimize parallel moves performed before a TailCall by grouping adjacent
> pushes. This way, we may use a single instruction to push multiple registers at
> once. However, we also have support for pushing immediates and stack slots for
> which the benefit is questionnable therefore this patch removes support for
> them.
>
> Concerning immediate pushes, it looks like a mistake since we do not have
> support for this case in `AssembleMove` so this patch removes it. Furthermore,
> if we add a test for this case, we see that a `push ip` instruction is
> generated, effectively pushing whatever was in `ip` at the time instead of
> pushing a constant.
>
> Concerning stack slot pushes, we generate a more or less equivalent sequence of
> instructions.
>
> Finally, grouping floating point pushes is not used anywhere so this patch
> removes support for this also.
>
> Bug: v8:6553
> Change-Id: I9b820d33361fc442dd813f66e1f96cda41009110
> Reviewed-on: https://chromium-review.googlesource.com/567191
> Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
> Commit-Queue: Pierre Langlois <pierre.langlois@arm.com>
> Cr-Commit-Position: refs/heads/master@{#46718}
R=pierre.langlois@arm.com, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=
LOG=N
Change-Id: I8790c7a72f92803ea8fda3c6dc7e6b013e2e09e9
Reviewed-on: https://chromium-review.googlesource.com/588471
Reviewed-by: Junliang Yan <jyan@ca.ibm.com>
Commit-Queue: Jaideep Bajwa <bjaideep@ca.ibm.com>
Cr-Commit-Position: refs/heads/master@{#46949}
The approach to creating compiler dispatcher jobs for inner functions after
they had been parsed didn't provide the startup benifits we hoped for due
to the need to hold onto the whole zone memory AST while waiting for the
jobs to complete.
This CL removes the ability to create these compilation jobs (which was never
enabled by default anyway). Going forward we will potentially use the
parser task approach to parse+compile inner functions in their own job.
BUG=v8:5203
Change-Id: I63134746aa23b4aa6e3bfa17c539954890fd9b0f
Reviewed-on: https://chromium-review.googlesource.com/590007
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46948}
This CL moves us much closer to the point where we can remove debugger-script.js and usage of debugger context from inspector.
There are three main parts left:
- managing breakpoints,
- inspecting stack and scopes (this CL),
- LiveEdit.
In this CL I moved all stack/scope inspection to native. As side effect running debugger and inspector tests are 10-20% faster (it's significant since not all of tests requesting break).
R=yangguo@chromium.org,jgruber@chromium.org
Bug: chromium:652939
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_chromium_rel_ng
Change-Id: I409396a687e18e9c0554c0c9c35b6e1064627be8
Reviewed-on: https://chromium-review.googlesource.com/580645
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46947}
This is preparation work to re-enable the scheduler: on arm64 some opcodes will
be neutral wrt the stack (JSSP) but will modify the underlying CSP. Identify
those opcode as such until JSSP is removed.
Bug:
Change-Id: Iae633382c5ed38b01edaec896f2ce44d76931fc8
Reviewed-on: https://chromium-review.googlesource.com/568822
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Rodolph Perfetta <rodolph.perfetta@arm.com>
Cr-Commit-Position: refs/heads/master@{#46946}
These counters handle cases in which the catch/finally block contains a jump
statement.
Bug: v8:6000
Change-Id: Ic83f11ee431edabe61f129c9abc3adc12a79c338
Reviewed-on: https://chromium-review.googlesource.com/586595
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46945}
We must ensure required root objects will be filled when such
objects will be deserialized.
Change-Id: I25136d31cb2e0c0a69a51c5635192f17bbe2a9ba
Reviewed-on: https://chromium-review.googlesource.com/579768
Commit-Queue: Vyacheslav Chigrin <vchigrin@yandex-team.ru>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46944}
We've been passing a context to the compiler, which turns out to be
solely used to determine if we're executing in a specific cctest configuration.
This change adds a configuration to the graph builder that we can use to
explicitly opt out of stack checks and traps. CcTests default to opting out,
except for the few that don't.
Bug:
Change-Id: I4724e31c2a62e9b3ab4feadb788287c374b39f53
Reviewed-on: https://chromium-review.googlesource.com/585779
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Mircea Trofin <mtrofin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46943}
This fixes the second-order Array.prototype function {forEach} and {map}
to now perform a callability check of the given callback function. For
empty arrays it is observable whether such a check outside the loop has
been elided or not.
R=mvstanton@chromium.org
TEST=mjsunit/regress/regress-crbug-747062
BUG=chromium:747062
Change-Id: I1bbe7f44b3b3d18e9b41ad0436975434adf84321
Reviewed-on: https://chromium-review.googlesource.com/588893
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46942}
This fixes a subtle register allocation verifier bug: depending on how moves are
optimized, and when having duplicate phis, we may end up overwriting a final
assessment.
Bug:
Change-Id: I8b7891efbdd075aae0219d60270f405b13d50f40
Reviewed-on: https://chromium-review.googlesource.com/577288
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Mircea Trofin <mtrofin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46941}
Reland of https://chromium-review.googlesource.com/c/558290/
Makes compiler dispatcher jobs an abstract interface, with unoptimized
compile jobs as an implementation of this interface.
Bug: v8:6537
Change-Id: Ia85781f72c7aaca497896ca4efa91ada97e43b1c
Reviewed-on: https://chromium-review.googlesource.com/589154
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46940}
In the process, cleanup some of the maths and functionality used to setup
descriptors and compute parameters. Also cleanup and correct the context
passing.
Change-Id: I6b6629bc81ef1c03425332dd6eadf3085efec7c9
Reviewed-on: https://chromium-review.googlesource.com/588892
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Daniel Clifford <danno@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46938}
Note that this also renames the existing "asm_wasm" variant to use the
more appropriate "stress_asm_wasm" name.
R=rmcilroy@chromium.org
BUG=v8:6409
Cq-Include-Trybots: master.tryserver.v8:v8_linux_noi18n_rel_ng
Change-Id: I1f9550cd03874c678f4583047a4e123a6f090250
Reviewed-on: https://chromium-review.googlesource.com/584879
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46937}
Instead of having feedback vector as a subtype of FixedArray with
reserved slots, make it a first-class variable-sized object with a
fixed-size header. This allows us to compress counters to ints in the
header, rather than forcing them to be Smis.
Cq-Include-Trybots: master.tryserver.chromium.linux:linux_chromium_rel_ng
Change-Id: Icc5f088ffbc2e2651b845bc71ea42060639e3e48
Reviewed-on: https://chromium-review.googlesource.com/585129
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46935}
Port 4e207a429a
Original Commit Message:
This switches the "code entry" field on JSFunction to no longer be an
inner pointer into a Code object (i.e. to the start of the instruction
stream), but a properly tagged pointer instead.
Motivation behind this is the ability to treat this field regularly as
part of escape analysis in the optimizing compiler. Also simplifies the
object visitation for JSFunction objects.
R=mstarzinger@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=
LOG=N
Change-Id: Ifa5998551e041c8de647df7306dd549455936699
Reviewed-on: https://chromium-review.googlesource.com/588468
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Jaideep Bajwa <bjaideep@ca.ibm.com>
Cr-Commit-Position: refs/heads/master@{#46934}
Test mjsunit/optimized-map walked an array through different
ElementsKind transitions, but it failed to verify that the
expected ElementsKind was in place. Although we have a regression
test for the bug, it's a good idea to make sure the basic
test covers all paths.
Bug: chromium:747075
Change-Id: I1424880801857f3356bfd63839d351d6fd1521e0
Reviewed-on: https://chromium-review.googlesource.com/584837
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Michael Stanton <mvstanton@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46933}
Makes compiler dispatcher jobs an abstract interface, with unoptimized
compile jobs as an implementation of this interface.
Bug: v8:6537
Change-Id: I6569060a89c92d35e4bc7962623f77082a354934
Reviewed-on: https://chromium-review.googlesource.com/558290
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46931}
The code generated for the TestTypeOf bytecode was not ideal, mostly
because of the default case that just aborted. If we do CSA_ASSERT to
check the validity of the literal_flag instead anf then just use the
last label as the default, the bytecode handler no longer builds a
stack frame and generated code quality is now really close to ideal.
The TestTypeOf bytecode handler was found to be among the three
hottest bytecode handlers in the Speedometer/AngularJS benchmark.
R=jarin@chromium.org
Change-Id: I47705a0ca0a436d5c42899001064e77d44845a64
Reviewed-on: https://chromium-review.googlesource.com/589207
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46930}
This replaces the linked list of weak cells with a worklist.
TBR=yangguo@chromium.org
BUG=chromium:694255
Change-Id: Ia877e25010ebbec9c05fbbe48cff460a92d3a132
Reviewed-on: https://chromium-review.googlesource.com/587067
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46929}
MSVC's macro VA_ARGS support behaves slightly differently from
gcc/clang, where VA_ARGS is treated as a single token when passed to
other macros, e.g.
#define FIRST(X, ...) 'X'
#define FOO(...) FIRST(__VA_ARGS__)
FOO(a,b,c)
expands to
gcc/clang: 'a'
MSVC: 'a,b,c'
The workaround to this is to wrap the call in a no-op macro, which
expands VA_ARGS first, and only then passes it through:
#define EXPAND(x) x
#define FOO(...) EXPAND(FIRST(__VA_ARGS__))
This was causing errors on windows builds when CSA_ASSERT was passed
multiple additional expressions.
Change-Id: Ia40bf23baf97af29c7f6f67c8a83918ecca15364
Reviewed-on: https://chromium-review.googlesource.com/586831
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46927}
This is a step in making shared function info visitor side effect free.
Change-Id: I5e7a713d2d782590c4988abc69b2068dc85b9bae
Reviewed-on: https://chromium-review.googlesource.com/586927
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46926}
Debug::ClearCoverageInfo may be called multiple times, for example
whenever coverage mode is switched to BestEffort. When encountering a
DebugInfo that does not have a CoverageInfo attached, simply do nothing
instead of failing a DCHECK.
TBR=yangguo@chromium.org
Bug: v8:6000
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_chromium_rel_ng
Change-Id: I3910b21ab40e7db02af5f3b6add3ec3f86e5bd91
Reviewed-on: https://chromium-review.googlesource.com/588788
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46925}
We need to write the slot with a relaxed write. No additional barrier is
needed because the new object has been published using Release_CAS and
reading either value is fine (as long as all initializing stores have
been published).
NOTRY=true
Bug: chromium:738865
Change-Id: Ieeacccf00e0a7e8eb950eee01060bd1963e30a42
Reviewed-on: https://chromium-review.googlesource.com/589007
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46924}
Adding (very) small tests for deoptimization.
Some of these tests were failing when the safepoints were not found,
after setting the return address.
BUG=V8:6563
Change-Id: I3af36b193a5982cd73414cc1884c5f0a7a727f5a
Reviewed-on: https://chromium-review.googlesource.com/584751
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Juliana Patricia Vicente Franco <jupvfranco@google.com>
Cr-Commit-Position: refs/heads/master@{#46922}
This reverts commit 862d605c13.
Reason for revert: fixed compile issue
Original change's description:
> Revert "[wasm] Consolidate function table representation."
>
> This reverts commit 4a45f35f26.
>
> Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/25471
>
> Original change's description:
> > [wasm] Consolidate function table representation.
> >
> > This CL avoids the need to reference the function tables (and signatures)
> > as either fixed arrays or vectors, preferring vectors.
> >
> > The only place we need fixed arrays is on the compiled module, to support
> > serialization. When we move off the GC heap, we'll also move away
> > from fixed arrays in that last case.
> >
> > The CL aids with getting wasm of the GC heap, by reducing the places
> > and representations we'll need to change when changing the way we
> > reference fixed tables.
> >
> > Bug:
> > Change-Id: Id4e43905a3df39062bf2839fa72dd5d9a0fe87da
> > Reviewed-on: https://chromium-review.googlesource.com/588334
> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org>
> > Reviewed-by: Brad Nelson <bradnelson@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#46917}
>
> TBR=bradnelson@chromium.org,titzer@chromium.org,mtrofin@chromium.org,ahaas@chromium.org
>
> Change-Id: Ie7d04f7ec74d6d0b3783df1c78c91c100ab784f4
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Reviewed-on: https://chromium-review.googlesource.com/588627
> Reviewed-by: Mircea Trofin <mtrofin@chromium.org>
> Commit-Queue: Mircea Trofin <mtrofin@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#46918}
TBR=bradnelson@chromium.org,titzer@chromium.org,mtrofin@chromium.org,ahaas@chromium.org
Change-Id: Ic0ba8097c13f2b1afd263b6243360e8ab95ae474
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/588667
Commit-Queue: Mircea Trofin <mtrofin@chromium.org>
Reviewed-by: Mircea Trofin <mtrofin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#46919}
