Introduce a new JSGlobalSpecialization advanced reducer that runs
during the initial inlining and context specialization, and specializes
the graph to the globals of the native context. Currently we assume
that we do not inline cross native context, but long-term we will grab
the global object from the JSLoadGlobal/JSStoreGlobal feedback (with the
new global load/store ICs that are currently in the workings), and then
this whole specialization will be fully compositional even across
cross-context inlining.
Note that we cannot really handle most of the stores to global object
property cells because TurboFan doesn't have a mechanism to enforce
certain representations. Also note that we cannot yet fully benefit
from the type feedback collected on the global object property cells,
because the type system cannot deal with maps in a reasonable way.
CQ_INCLUDE_TRYBOTS=tryserver.v8:v8_linux_nosnap_rel
R=jarin@chromium.org
BUG=v8:4470
LOG=n
Committed: https://crrev.com/6fbf7903f94924ea066af481719898bd9667b6eb
Cr-Commit-Position: refs/heads/master@{#31139}
Review URL: https://codereview.chromium.org/1387393002
Cr-Commit-Position: refs/heads/master@{#31148}
Reason for revert:
Breaks GC stress: http://build.chromium.org/p/client.v8/builders/V8%20Linux64%20GC%20Stress%20-%20custom%20snapshot/builds/1984/steps/Bisect%20c5528ac1.Retry/logs/regress-crbug-450960
Original issue's description:
> [turbofan] Add initial support for global specialization.
>
> Introduce a new JSGlobalSpecialization advanced reducer that runs
> during the initial inlining and context specialization, and specializes
> the graph to the globals of the native context. Currently we assume
> that we do not inline cross native context, but long-term we will grab
> the global object from the JSLoadGlobal/JSStoreGlobal feedback (with the
> new global load/store ICs that are currently in the workings), and then
> this whole specialization will be fully compositional even across
> cross-context inlining.
>
> Note that we cannot really handle most of the stores to global object
> property cells because TurboFan doesn't have a mechanism to enforce
> certain representations. Also note that we cannot yet fully benefit
> from the type feedback collected on the global object property cells,
> because the type system cannot deal with maps in a reasonable way.
>
> CQ_INCLUDE_TRYBOTS=tryserver.v8:v8_linux_nosnap_rel
> R=jarin@chromium.org
> BUG=v8:4470
> LOG=n
>
> Committed: https://crrev.com/6fbf7903f94924ea066af481719898bd9667b6eb
> Cr-Commit-Position: refs/heads/master@{#31139}
TBR=jarin@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:4470
Review URL: https://codereview.chromium.org/1390073004
Cr-Commit-Position: refs/heads/master@{#31144}
Thus TypeFeedbackMetadata can now be shared between different native contexts.
Review URL: https://codereview.chromium.org/1384673002
Cr-Commit-Position: refs/heads/master@{#31143}
Adds support for compiling top level code to bytecode to be run in the
interpreter.
Also moves PassesFilter to String:: so that it can be used to filter top
level script names as well as functions (used in
https://codereview.chromium.org/1379093002/)
BUG=v8:4280
LOG=N
Review URL: https://codereview.chromium.org/1372293005
Cr-Commit-Position: refs/heads/master@{#31142}
Introduce a new JSGlobalSpecialization advanced reducer that runs
during the initial inlining and context specialization, and specializes
the graph to the globals of the native context. Currently we assume
that we do not inline cross native context, but long-term we will grab
the global object from the JSLoadGlobal/JSStoreGlobal feedback (with the
new global load/store ICs that are currently in the workings), and then
this whole specialization will be fully compositional even across
cross-context inlining.
Note that we cannot really handle most of the stores to global object
property cells because TurboFan doesn't have a mechanism to enforce
certain representations. Also note that we cannot yet fully benefit
from the type feedback collected on the global object property cells,
because the type system cannot deal with maps in a reasonable way.
CQ_INCLUDE_TRYBOTS=tryserver.v8:v8_linux_nosnap_rel
R=jarin@chromium.org
BUG=v8:4470
LOG=n
Review URL: https://codereview.chromium.org/1387393002
Cr-Commit-Position: refs/heads/master@{#31139}
It was supposed to be used by the CPU profiler. But as long as
these ranges are not built when profiler is not running, once
the profiler is started there're no ranges for already compiled
functions. So basically this code never worked.
As long as now CPU profiler uses another approach this code is no
longer needed.
Review URL: https://codereview.chromium.org/1376333003
Cr-Commit-Position: refs/heads/master@{#31056}
This CL also allows to use arbitrary number of feedback vector elements for particular slot kind.
Review URL: https://codereview.chromium.org/1370303004
Cr-Commit-Position: refs/heads/master@{#31050}
The LiteralsArray will soon hold a type feedback vector. Code treats it as an
ordinary fixed array, and needs to stop that.
BUG=
Review URL: https://codereview.chromium.org/1374723002
Cr-Commit-Position: refs/heads/master@{#31000}
Replacing it with SMI_ACCESSORS.
This change makes accesses to Smi fields in objects more regular (the
accessors now always consume/return an int rather than a Smi*), which
avoids a bunch of manual Smi::FromInt() and Smi::value() conversions,
and is a step on the way towards being able to generate objects-inl.h.
Review URL: https://codereview.chromium.org/1371893002
Cr-Commit-Position: refs/heads/master@{#30975}
This name makes it clear that the flag (also the variant in the Compiler)
is talking about specializing to the function context instead of i.e. the
native context.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/1372513003
Cr-Commit-Position: refs/heads/master@{#30934}
We're moving away from using CompilationInfo as a big bag o' stuff.
Passing in just what we need to several AstVisitors to avoid
increasing the problem.
BUG=None
TEST=trybots
R=titzer@chromium.org
LOG=N
Review URL: https://codereview.chromium.org/1318823010
Cr-Commit-Position: refs/heads/master@{#30529}
This CL us a pure refactoring that makes an empty compilation unit
including just "isolate.h" or "contexts.h" but not "objects-inl.h"
compile without warnings or errors. This is needed to further reduce
the header dependency tangle.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/1322883002
Cr-Commit-Position: refs/heads/master@{#30500}
Replaces all instances of the code which computed the debug
name of a stub or function with calls to CompileInfo::GetDebugName instead.
Also:
- Removes useless parameter on CodeStub::GetMajorName
- Removes FakeStubForTesting since it is no longer required
- Adds CompileInfo::ShouldEnsureSpaceForLazyDeopt() to replace unclear calls to IsStub().
Review URL: https://codereview.chromium.org/1297203002
Cr-Commit-Position: refs/heads/master@{#30324}
While the class Type objects the AstTyper generates should be
allocated to zone memory, there's no particular reason
the typer object itself needs to be.
BUG=None
TEST=None
R=rossberg@chromium.org,titzer@chromium.org
LOG=N
Review URL: https://codereview.chromium.org/1303843003
Cr-Commit-Position: refs/heads/master@{#30313}
We can use the script type to determine that instead. Script of type
TYPE_NATIVE are considered builtins, TYPE_NORMAL are not. The only exception
to this rule is the empty function, for which the script is TYPE_NATIVE
(observable by the debugger), but should be stringified to "function () {}"
instead of "function () { [native code] }". For this, I introduce a
hide_source flag on the script object.
We also use IsBuiltin and IsSubjectToDebugging interchangeably. For debugger,
we now use the latter, hiding the detail that only non-builtins are debuggable.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/1292283004
Cr-Commit-Position: refs/heads/master@{#30285}
Bytecode generator for local assignment and basic binary operations.
Command-line flag for printing bytecodes.
BUG=v8:4280
LOG=N
Review URL: https://codereview.chromium.org/1294543002
Cr-Commit-Position: refs/heads/master@{#30221}
This is the first step of turning the v8.h file into a normal header
instead of an include-the-world header. The new rule is that no other
header files are allowed to include v8.h, which is enforced by DEPS.
Also the number of includes inside the v8.h file has been drastically
reduced. Basically the last missing piece is the inclusion of the big
objects-inl.h file.
This in turn makes many headers follow the IWYU principle.
R=bmeurer@chromium.org,hpayer@chromium.org,titzer@chromium.org
Review URL: https://codereview.chromium.org/1282503003
Cr-Commit-Position: refs/heads/master@{#30102}
This also allows us to not always compile for debugging when debug is active.
Review URL: https://codereview.chromium.org/1258583002
Cr-Commit-Position: refs/heads/master@{#29848}
The background parser checks for debugger state in its constructor. This
is not good enough, since the debugger state may change afterwards, but
before compiling takes place. As the background parser can only parse
lazily, this could mean that due to debugging, we try to eagerly compile
an inner function we have not eagerly parsed.
R=jochen@chromium.org
Review URL: https://codereview.chromium.org/1247743002
Cr-Commit-Position: refs/heads/master@{#29784}
Prior to this patch, we enter a global debug mode whenever a break point
is set. By entering this mode, all code is deoptimized and activated
frames are recompiled and redirected to newly compiled debug code.
After this patch, we only deoptimize/redirect for functions we want to
debug. Trigger for this is Debug::EnsureDebugInfo, and having DebugInfo
object attached to the SFI prevents optimization/inlining.
The result is that we can have optimized code for functions without break
points alongside functions that do have break points, which are not
optimized.
R=mstarzinger@chromium.org, ulan@chromium.org
BUG=v8:4132
LOG=Y
Review URL: https://codereview.chromium.org/1233073005
Cr-Commit-Position: refs/heads/master@{#29758}
EnumtSet has been deprecated for quite some time, and replaced with the
more general and type safe base::Flags template class.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1229233002
Cr-Commit-Position: refs/heads/master@{#29564}
We have to reland these two commits at once, because the first breaks
some asm.js benchmarks without the second. The change was reverted
because of bogus checks in the verifier, which will not work in the
presence of OSR (and where hidden because of the type back propagation
hack in OSR so far). Original messages are below:
[turbofan] Add new JSFrameSpecialization reducer.
The JSFrameSpecialization specializes an OSR graph to the current
unoptimized frame on which we will perform the on-stack replacement.
This is used for asm.js functions, where we cannot reuse the OSR
code object anyway because of context specialization, and so we could as
well specialize to the max instead.
It works by replacing all OsrValues in the graph with their values
in the JavaScriptFrame.
The idea is that using this trick we get better performance without
doing the unsound backpropagation of types to OsrValues later. This
is the first step towards fixing OSR for TurboFan.
[turbofan] Perform OSR deconstruction early and remove type propagation.
This way we don't have to deal with dead pre-OSR code in the graph
and risk optimizing the wrong code, especially we don't make
optimistic assumptions in the dead code that leaks into the OSR code
(i.e. deopt guards are in dead code, but the types propagate to OSR
code via the OsrValue type back propagation).
BUG=v8:4273
LOG=n
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/1226673005
Cr-Commit-Position: refs/heads/master@{#29486}
The JSFrameSpecialization specializes an OSR graph to the current
unoptimized frame on which we will perform the on-stack replacement.
This is used for asm.js functions, where we cannot reuse the OSR code
object anyway because of context specialization, and so we could as well
specialize to the max instead.
It works by replacing all OsrValues in the graph with their values in
the JavaScriptFrame.
The idea is that using this trick we get better performance without
doing the unsound backpropagation of types to OsrValues later. This is
the first step towards fixing OSR for TurboFan.
R=jarin@chromium.org
BUG=v8:4273
LOG=n
Review URL: https://codereview.chromium.org/1225683004
Cr-Commit-Position: refs/heads/master@{#29476}
This makes new.target work in [[Call]] and [[Construct]] of ordinary
functions.
We achieve this by introducing a new construct stub for functions that
uses the new.target variable. The construct stub pushes the original
constructor just above the receiver in the construct frame.
BUG=v8:3887
LOG=N
R=adamk@chromium.org, dslomov@chromium.org
Review URL: https://codereview.chromium.org/1203813002
Cr-Commit-Position: refs/heads/master@{#29358}
Now that we keep tabs on shared function infos from a script, we can speed up finding shared function infos for debugging. However, in case we have to compile a function that cannot be lazily compiled without context, we fall back to the slow heap iteration.
R=mstarzinger@chromium.org
BUG=v8:4132,v8:4052
LOG=N
Committed: https://crrev.com/cfe89a71a332ef9ed481c8210bc3ad6d2822034b
Cr-Commit-Position: refs/heads/master@{#29296}
Review URL: https://codereview.chromium.org/1206573004
Cr-Commit-Position: refs/heads/master@{#29327}
