CloneContext is implemented as a runtime call so it is considerably slower
than AllocateContext which leads to performance issues in loops that capture
iteration variables in for loops.
Lowering CloneContext into AllocateUninitializedContext and a sequence of
explicit load&store instructions both improves the performance by avoiding
the runtime call and also makes CloneContext semantics exposed to store-load
forwarding and allocation sinking.
PutIfAbsent benchmark from issue #31113 is improved by 5x, while
ForInGeneratedLoop benchmark is improved by 8x.
Bug: https://github.com/dart-lang/sdk/issues/31113
Change-Id: I03fdede75f1a85fda92487aa8f9307ce06f23903
Reviewed-on: https://dart-review.googlesource.com/16280
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This reverts commit ade37f931e.
This re-lands RebindRule change with fixes for the "-mdebug--hot-reload" test failures.
Original commit is reverted in patch set 1.
Fixes are in following patch sets.
Bug:
Change-Id: I49375af9b891323fe05c670d77cbf880964aae54
Reviewed-on: https://dart-review.googlesource.com/9361
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Alexander Aprelev <aam@google.com>
Change-Id: I6219462fed278dbb90361c2084eed886c9245097
Revert "Revert "Introduce inline cache reload rule attribute to static call AST node and ICData.""
This reverts commit 162283c2f3.
This re-lands the change with original revert in Patch set 1, actual bug fix in next patch set.
Bug:
Change-Id: I6219462fed278dbb90361c2084eed886c9245097
Reviewed-on: https://dart-review.googlesource.com/8760
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
This reverts commit 68a727fa5c.
Revert "Fix ICData::New() invocation. Follow-up to 68a727fa5cf5."
This reverts commit ac06b1c8ea.
Revert "Fix last ICData::New() invocation. Follow-up to 68a727fa5cf5."
This reverts commit da26ffc4de.
Bug:
Change-Id: Ia2d4e6a3c3a3a329982c90aded11e72bf25a9a34
Reviewed-on: https://dart-review.googlesource.com/8402
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Alexander Aprelev <aam@google.com>
Reload rule describes how inline cache has to be processed during hot reload, whether it should
be updated, reset or preserved.
Bug: dartbug.com/30639
Change-Id: I40b63ade786456ec48d4fa205a076654c2996bce
Reviewed-on: https://dart-review.googlesource.com/7586
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Do this in unoptimized code only, when --reify-generic-functions is specified.
This is still work in progress, and support in optimizer, in inliner, in DBC,
in kernel to ir, and other areas, will follow.
Many small fixes and added todos.
R=rmacnak@google.com, vegorov@google.com
Review-Url: https://codereview.chromium.org/2941643002 .
Original CL: https://codereview.chromium.org/2692803006/
Original commit message:
Tracking the awaiter return call stack:
- [x] Each async function closure now knows who is awaiting on their
return. This is effectively the asynchronous equivalent of the 'frame pointer'.
- [x] Each async* function closure now knows how is listening on their
stream. This is effectively the asynchronous equivalent of the 'frame pointer'.
Detecting uncaught exceptions in async functions:
- [x] Code object keeps a map from :await_jump_var to token position
- [x] Exception Handlers keep track if they are generated (as part of compilation) or directly from user code
- [x] Debugger maps :await_jump_var to a specific try index
R=johnmccutchan@google.com
Review-Url: https://codereview.chromium.org/2725623003 .
Tracking the awaiter return call stack:
- [x] Each async function closure now knows who is awaiting on their
return. This is effectively the asynchronous equivalent of the 'frame pointer'.
- [x] Each async* function closure now knows how is listening on their
stream. This is effectively the asynchronous equivalent of the 'frame pointer'.
Detecting uncaught exceptions in async functions:
- [x] Code object keeps a map from :await_jump_var to token position
- [x] Exception Handlers keep track if they are generated (as part of compilation) or directly from user code
- [x] Debugger maps :await_jump_var to a specific try index
Fixes#27242R=rmacnak@google.com
Review-Url: https://codereview.chromium.org/2692803006 .
When referring to constants via a deferred prefix in an await expression
the transformation must make sure that the generated code before and after
the deferred load agree on the number of captured variables. (i.e. the number
of await-temp variables introduced by the await-transformer is the same)
This CL uses a temporary in the case of compile-time constants, because before
the deferred load, a reference lib.C is translated into a static getter which
also requires a temporary.
Fixes#28678R=hausner@google.com
Review-Url: https://codereview.chromium.org/2683973002 .
In some cases when loading a deferred library twice, the VM would crash/throw.
Make sure we generate the correct static NoSuchMethod error when referring to
a deferred library that is not loaded yet.
Fixes#28278.
R=rmacnak@google.com
Review-Url: https://codereview.chromium.org/2618823002 .
i.e. #ifndef VM_WHATEVER -> #ifndef RUNTIME_VM_WHATEVER
This lets us remove a hack from the PRESUBMIT.py script that existed
for reasons that are no longer valid, and sets us up to add some
presubmit checks for the GN build.
R=asiva@google.com, rmacnak@google.com
Review URL: https://codereview.chromium.org/2450713004 .
Consider code like
try {
something();
} finally {
assert(expr);
}
The compiler de-sugars this into
try {
something();
} catch (_) {
rethrow;
} finally {
assert(expr);
}
In unchecked mode this leaves an empty finally-block, but we would still
have the overhead of the try-catch. This CL avoids this unnecessary overhead
by making try-finally with an empty finally zero-cost.
BUG=#27274
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/2322853003 .
We can have stores inlined in the optimizer that are not initializing stores.
Mark initializing stores in the parser instead to distinguish
initializing- and non-initializing stores more precisely.
BUG=
R=vegorov@google.com
Review URL: https://codereview.chromium.org/2112973002 .
Fourth(!) attempt. This CL fixes another instance where parsing a nested function modifies the parser state of the function that is being compiled.
When a local function gets compiled the second time, constant
expressions may not be parsed again, since the constant value
is found in the cache. If the expression refers to an outer
variable, it does not get captured correctly.
Fix: instead of parsing a local function repeatedly to capture
outer variables, use the local function’s context scope to mark
outer variables as captured. This fixes the bug, and makes the
compiler more efficient as well.
BUG= 26453
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/2010283004 .
This reverts commit 4dca5d0e01.
It seems to cause at least one additional test failure:
Repro:
python tools/test.py -mrelease -asimarm64 --write-debug-log --write-test-outcome-log --copy-coredumps -cprecompiler -rdart_precompiled --exclude-suite pkg -t480 language/async_star_regression_fisk_test
BUG=
R=vegorov@google.com
Review URL: https://codereview.chromium.org/2004373004 .
Third attempt. A latent bug in finally block inlining
caused the previous crashes in optimized functions.
Fix is in https://codereview.chromium.org/2004883004/.
When a local function gets compiled the second time, constant
expressions may not be parsed again, since the constant value
is found in the cache. If the expression refers to an outer
variable, it does not get captured correctly.
Fix: instead of parsing a local function repeatedly to capture
outer variables, use the local function’s context scope to mark
outer variables as captured. This fixes the bug, and makes the
compiler more efficient as well.
BUG=26453
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/2008043002 .
Second attempt, this time also capturing hidden variables
like 'this' in initializer lists. Original CL is
https://codereview.chromium.org/1980193002
When a local function gets compiled the second time, constant
expressions may not be parsed again, since the constant value
is found in the cache. If the expression refers to an outer
variable, it does not get captured correctly.
Fix: instead of parsing a local function repeatedly to capture
outer variables, use the local function’s context scope to mark
outer variables as captured. This fixes the bug, and makes the
compiler more efficient as well.
BUG=26453
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/1986393002 .
When a local function gets compiled the second time, constant
expressions may not be parsed again, since the constant value
is found in the cache. If the expression refers to an outer
variable, it does not get captured correctly.
Fix: instead of parsing a local function repeatedly to capture
outer variables, use the local function’s context scope to mark
outer variables as captured. This fixes the bug, and makes the
compiler more efficient as well.
No wall-time improvement found when running dart2js, though.
BUG=26453
R=regis@google.com
Review URL: https://codereview.chromium.org/1980193002 .
- Use TokenDescriptor instead of intptr_t for all token positions.
- Use TokenDescriptor in raw_object instead of int32_t.
- TokenDescriptor is a POD with an int32_t (this shrinks the size of AST and IR nodes by 32-bits on 64-bit architectures).
There are some cleanups I plan on doing as a follow up CL:
- Replace TokenDescriptor::value() with TokenDescriptor::TokenPos()
R=iposva@google.com, rmacnak@google.com
Review URL: https://codereview.chromium.org/1644793002 .
- Rework token position address space.
- Use ClassifyingTokenPositions::kMethodExtractor for the token position of a method extractor.
- Plumb token positions through all of Ast Transformer.
- Plumb token positions for temporary expressions, etc in Flow Graph Builder.
- Add token positions for parts of the await machinery.
- Move ClassifyingTokenPositions into token.h.
- Remove default token position of Scanner::kNoSourcePos for many IR instructions.
- A couple of unit tests.
- Use synthetic token positions for synthetic AstNodes.
- Fix SLEB128 encoding / decoding + test
- s/Scanner::kNoSourcePos/Token::kNoSourcePos.
- Remove >= 0 and < 0 checks against token positions and use helpers instead.
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/1589643002 .
- Give AwaitMarkerNode a token position.
- Make it possible to provide a token position for SaveContext, RestoreContext, StoreContext, and CurrentContext.
- Pass in a token position to internally generated StoreLocal and LoadLocal instructions for try-catch-finally.
- Introduce a Context ClassifyingTokenPosition to be used when we can't attribute a context related instruction to a source position.
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/1575953005 .
There were two exception handlers with the same index added to the code so that only the one added
last was executed. In case of an exception that means that a re-throw may be omitted, causing invalid
control flow.
Also, add assertion to ensure unique try-index for exception handlers.
BUG=#25333
R=hausner@google.com
Review URL: https://codereview.chromium.org/1569523002 .
- Rename accessors of class Field to make it more apparent as to what is being accessed (static field values or instance field offsets
- Use precompiled initializer state (Used only during precompilation) to also store saved initial value of static fields (Used only during application snapshot generation)
BUG=
R=hausner@google.com, rmacnak@google.com
Review URL: https://codereview.chromium.org//1289643005 .
The first invocation of a native functions goes into LinkNativeCall which
determines the final entry point and patches the object pool entry.
When running precompiled code, this makes deserializing the object
pool entries for native functions easy, they all initially point to
a single entry (LinkNativeCall).
BUG=
R=rmacnak@google.com
Review URL: https://codereview.chromium.org//1294113004 .
The new flag --enable-debug-break turns the otherwise illegal Dart statement
break "message";
into a break instruction preceded with a debug message, the equivalent of
emitting an Assembler::Stop("message").
Add a language test expecting a syntax error without the flag.
Change expected break instruction in arm64 simulator.
Remove constants related to now deleted simulator tracing on mips and arm64.
R=johnmccutchan@google.com, srdjan@google.com
Review URL: https://codereview.chromium.org//1087383002
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@45218 260f80e4-7a28-3924-810f-c04153c831b5
Wrap the await-for loop in a try-finally statement that ensures that
the stream is cancelled. Also adding the ability to the
AsyncStartStreamController class to return a cancellation future.
The cancel() call at the end of the await-for loop is not awaiting
the cancellation future yet. This is not part of the spec at this
point.
R=regis@google.com
Review URL: https://codereview.chromium.org//1014273003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@44689 260f80e4-7a28-3924-810f-c04153c831b5
Vyacheslav Egorov