Added a separate flag for this, since we intend to enable it for the linear allocator as well. Currently, the option is "on" for greedy, as a point in time to enable its testing (through the greedy allocator bots).
BUG=
Review URL: https://codereview.chromium.org/1256313003
Cr-Commit-Position: refs/heads/master@{#30005}
The race occurred when Workers were used. Since Workers call
Shell::ExecuteString from a different thread, TSAN (correctly) flags
this as a racy write. Solution would be to either synchronize the writes,
or to 'lift' the write higher up in the call stack and only write the flag
from the main thread. This implements this latter solution.
These methods call Shell::ExecuteString, but do *not* set script_executed:
- ExecuteInThread: Can only occur is JS has already been executed.
- Shell::Load: Callback for JS; so JS has already been executed when
we get there.
- Shell::RunShell: Interactive shell. We no longer need script_executed once
we're here.
BUG=v8:4330
LOG=N
Review URL: https://codereview.chromium.org/1258303004
Cr-Commit-Position: refs/heads/master@{#30003}
Store arity in FormalParameters; store name (instead of var) and is_rest flag in individual parameters. Ensure that the arity is always maintained consistently.
This is preparation for more parameter destructuring adjustments. In particular, a follow-up CL will separate parameter recording from declaring the variables.
R=adamk@chromium.org, littledan@chromium.org
BUG=v8:811
LOG=N
Review URL: https://codereview.chromium.org/1259013003
Cr-Commit-Position: refs/heads/master@{#30002}
They need to be properly recorded in the scope's temps set, otherwise allocation doesn't know about them and can break. (Not observable right now, but necessary for follow-up changes to parameter destructuring.)
Also, print temporary variables in a useful manner.
R=adamk@chromium.org
BUG=
Review URL: https://codereview.chromium.org/1263563002
Cr-Commit-Position: refs/heads/master@{#29998}
The heuristic can cause weird behavior when bootstrapping.
The memory savings is not worth this hassle.
Review URL: https://codereview.chromium.org/1265983006
Cr-Commit-Position: refs/heads/master@{#29992}
That makes going to the page safe. Addresses can be in arbitrary locations of an object, e.g. in a large object but not on the first 1M page.
BUG=
Review URL: https://codereview.chromium.org/1256203004
Cr-Commit-Position: refs/heads/master@{#29991}
Change minimum BytecodeArray frame size to zero now return value is in
the accumulator.
Fix inconsistent checks in bytecode-array-builder.cc.
Simplify bytecode disassembly by adding Bytecodes::Decode to
disassemble one bytecode and operands.
BUG=v8:4280
LOG=N
Review URL: https://codereview.chromium.org/1259193004
Cr-Commit-Position: refs/heads/master@{#29988}
There is only one use case for it: String.prototype.search converts a
string argument into a RegExp. The cache is used to avoid repeating that
conversion. However, this does not make the added complexity worthwhile.
Review URL: https://codereview.chromium.org/1267493006
Cr-Commit-Position: refs/heads/master@{#29985}
High promotion mode limits new space size in the case where many objects (high
rate) are promoted. The result is frequent scavenger runs with low pause times. The
approach conflicts with pretenuring which requires a larger new space size to
collect enough feedback to be usable.
Ultimately the goal is to have pretenuring catch allocation sites that have high
promotion rates.
Revert "Bring back high promotion mode to shrink young generation size when scavenging latency is high."
This reverts commit 37be1d5e9e.
Revert "GC. Delay/avoid entering high promotion mode"
This reverts commit a5221d0779.
BUG=chromium:499507,chromium:506910
LOG=n
Review URL: https://codereview.chromium.org/1247723003
Cr-Commit-Position: refs/heads/master@{#29983}
Reason for revert:
Reverted because 507840 came back on recent Chromecrash. Should not have committed this Cl.
Original issue's description:
> Reland concurrent sweeping of code space.
>
> BUG=
>
> Committed: https://crrev.com/8516dccf6a561020441773c93c564dd4aa6ee59e
> Cr-Commit-Position: refs/heads/master@{#29967}
TBR=jochen@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/1263343002
Cr-Commit-Position: refs/heads/master@{#29975}
The calculation now takes into account the size of the arguments object
if it is present in the optimized frame.
(Yang, many thanks for the awesome repro!)
BUG=chromium:514362
LOG=N
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/1264483008
Cr-Commit-Position: refs/heads/master@{#29973}
This fixes a bug introduced by r28826 (Unify decoding of deoptimization
translations, https://codereview.chromium.org/1136223004), where we
started leaking arguments marker sentinel to the debugger, which would
then cause crashes. This change replaces the sentinel with the undefined
value in the debugger-inspectable frame.
BUG=chromium:514362
LOG=n
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/1263333002
Cr-Commit-Position: refs/heads/master@{#29971}
The BytecodeArrayBuilder has responsibility for emitting the BytecodeArray. It will be used by the AST walker.
Bytecode now uses an accumulator plus registers rather being pure register based.
Update BytecodeArray::Disassemble to print operand information.
BUG=v8:4280
LOG=N
Review URL: https://codereview.chromium.org/1266713004
Cr-Commit-Position: refs/heads/master@{#29970}
port ec9bc79473 (r29949).
original commit message:
Previously these instructions tried to jump to the value at the code entry's
location, rather than jumping to this location. Also adds a test.
BUG=
Review URL: https://codereview.chromium.org/1256163003
Cr-Commit-Position: refs/heads/master@{#29964}
port 1a5751f9b3 (r29956)
original commit message:
Since we need the notion of a dummy vector ic, we can use that to avoid
a special case of the IC constructor. Also, consolidate the two dummy
ICs into one.
BUG=
Review URL: https://codereview.chromium.org/1265113002
Cr-Commit-Position: refs/heads/master@{#29963}
Introduce new mechanism for relocating j/jal.
Resolves flaky failures of mozilla regress tests.
Additionally:
- internal encoded references are not relocated during code generation phase.
- remove asserts from j and jal which are not
valid because addresses are not final and valid in code generation phase.
TEST=mozilla/js1_5/Regress/regress-280769-2, regress-367561-01,
mozilla/ecma_3/Statements/regress-444979
BUG=
R=paul.lind@imgtec.com
Review URL: https://codereview.chromium.org/1216823003 .
Patch from dusan.milosavljevic <dusan.milosavljevic@imgtec.com>.
Cr-Commit-Position: refs/heads/master@{#29962}