Change the interface to the Hydrogen graph builder to appear like it
directly holds a current basic block and a current environment.
Remove some direct accesses to the current subgraph, and remove
subgraph accessors that simply forwarded to the exit block.
Review URL: http://codereview.chromium.org/6532088
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6887 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Before, the live ranges of the arguments extended to the call itself, and
they were pushed immediately before the call. Now, they are spilled eagerly
as soon as their value is available and they are spilled to the right place.
The inlined runtime calls in the optimized backend are changed to work as in
all the other backends: they get their arguments untranslated and can choose
their own custom evaluation order.
Review URL: http://codereview.chromium.org/6526047
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6876 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The baseline compiler does not materialize a value for expressions of
the form !expr in an effect context so the graph translation should
not produce such an environment, otherwise we risk targeting it by
deoptimization.
BUG=v8:1167
Review URL: http://codereview.chromium.org/6537018
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6833 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When deoptimizing to after an expression of the form (expr0 || expr1)
or (expr0 && expr1) in an effect context, the unoptimized code could
incorrectly see the value of the expression.
Handle the short-circuit binary operators specially in effect contexts.
This fixes the issue and will generate better code when the left
subexpression is boolean-valued.
BUG=v8:1166
Review URL: http://codereview.chromium.org/6519046
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6831 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Rather than passing in a pair of unequal-length lists, pass the default
subgraph separately. Construct the typecase from the top down rather than
the bottom up, so it doesn't need an intermediate zone list.
Also, change a basic block's 'last' instruction field to really be its last
instruction by correctly updating it when inserting and removing
instructions.
Review URL: http://codereview.chromium.org/6516016
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6783 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Before, Hydrogen call instructions had uses of the PushArgument instructions
for their arguments. These operands were unneeded, bloated the IR, and
caused calls to be the only Hydrogen instructions with an unpredictable
number of operands.
Now, PushArgument is a pure side-effecting instruction that has no uses.
Review URL: http://codereview.chromium.org/6480030
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6749 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Before, an attempt to delete a parameter in a function that used the
arguments object in any way would succeed with true and delete both
the parameter and the corresponding arguments object property.
Now, an attempt to delete such a parameter does not delete and
evaluates to false.
Parameters can be deleted, as before, from functions that use the
arguments object, by deleting the corresponding arguments object
property (this is a spec violation).
BUG=fixes v8:1136
Review URL: http://codereview.chromium.org/6484023
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6745 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Make esi available to the register allocator rather than dedicating it
permanently to the context.
The context is still passed in register esi to JavaScript and to the runtime
as part of the calling convention. Because some stubs might end up calling
JS or the runtime, it is also conservatively passed to stubs.
Roughly half the calls have been modified to use the context as an input
value in fixed register esi. The other half are marked as calls or deferred
code so esi is spilled and can be explicitly set.
It is no longer necessary to restore the context to esi after a call that
might change it.
Review URL: http://codereview.chromium.org/6452001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6713 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When constructing stack traces we interpret the deoptimization data for
optimized frames to find the receiver value. This value could sometimes be
eliminated from the deoptimization data if we though it was unused.
BUG=v8:1118
Review URL: http://codereview.chromium.org/6465023
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6694 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Record a safepoint with a deoptimization id for throw in optimized code. We
don't seem to much care what the AST ID is because we will not be using it
for lazy deoptimization (throw doesn't return to the point of throw). For
hygiene we use the actual ID of the throw expression. Throw is no longer a
control-flow instruction, but it's followed by an unconditional abnormal
exit. This is required to insert a simulate between the throw and the exit.
Make our optimized treatment of Function.prototype.apply act like a call and
have side effects. This ensures that it will get a lazy deoptimization
environment. Use that deoptimization ID in the safepoint for the call.
Deleting a property was also missing a deoptimization ID, though there was a
deoptimization environment assigned to the instruction. Record the
environment and use the deoptimization ID at the safepoint.
Review URL: http://codereview.chromium.org/6250105
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6576 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
1. The placement of checks for negative zero has to be computed after
all conversion instructions have been inserted. I separated the code
into its own phase.
2. GVN need to take instruction flags into account when comparing
instructions for redundancy.
Review URL: http://codereview.chromium.org/6260035
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6534 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Deletion of global properties puts 'the hole' in the global property
cell and updates the property details in the property dictionary with
the information that the property has been deleted. When setting
global properties that have been deleted in generated code we just
store the new value in the global property cell. This does not update
the property details in the property dictionary. Therefore, it looks
like the property is not there eventhough it was just reintroduced.
Perform 'the hole' checks in generated code for global property stores
and bail out of ICs and optimized code if storing to a property cell
that contains 'the hole'.
Review URL: http://codereview.chromium.org/6306014
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6508 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Before, we disabled code motion for check instructions on the final
optimization attempt. It is unsafe to prevent movement of check
instructions but to allow movement of instructions that assume the checks
were performed.
Review URL: http://codereview.chromium.org/6378012
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6485 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
To avoid deopts a few extra changes were needed:
o Enable megamorphic state for special property loads on
primitives. We used to flip between monomorphic stubs.
o Extract pure string (no string wrapper support) version of the
string length stub.
Review URL: http://codereview.chromium.org/6334015
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6472 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Rename HBranch (the instruction that coerces an arbitrary HValue to
control flow) to HTest to free up the term Branch to refer to any
control instruction with two successors.
Change the virtual FirstSuccessor and SecondSuccessor functions on
control instructions to a pair of data members.
Review URL: http://codereview.chromium.org/6366002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6417 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This patch adds H- and L-variants of StringCharCodeAt and StringLength.
StringCharCodeAt is used to inline a constant function call of
String.prototype.charCodeAt and to implement the corresponding inline
runtime function. It does not yet use the recently introduced extra IC
state. (We can specialize on string encoding and avoid deopts because
of out of bounds accesses.)
StringLength needs more work because the stub version of it also
supports strings wrappers and it matters in some cases. (We have to
separate the string only case.)
Review URL: http://codereview.chromium.org/6243008
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6408 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This instruction only depends on the prototype and the holder and can
completely ignore the receiver and its map.
This change also fixes a small bug on arm where a cell was loaded
instead of the prototype from new space.
Review URL: http://codereview.chromium.org/6094020
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6290 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Each LInstruction is now a subclass of LTemplateInstruction<R, I, T>
where R is number of outputs, I number of inputs and T number of temps.
This change only actually uses the parameter I for input operands.
Since the parameter T for temps is 0, it incurs no extra cost.
A separate change will introduce using the temps parameter.
Review URL: http://codereview.chromium.org/6215002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
For polymorphic calls (also loads and stores) we construct a type switch
graph that has a basic block merging all the variants. There is an
environment simulation before the goto at the end of all the predecessor
blocks. This simulation is used to define the environment on entry to the
successor block, and captures the return value of the call. In effect
contexts, this value should not be present in the environment.
The fix is to use the AST context to decide whether to have this value in
the join node's environment at all.
BUG=1014
Review URL: http://codereview.chromium.org/6065014
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6189 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
If the instance of is performed against what is beliwed to be a constant global function inline the instance of check and have the call to the instanceof stub in deferred code. The inlined check will be patched by the instanceof stub when called from deferred code. This is indicated by the lithium instruction LInstanceOfKnownGlobal.
To help the patching the delta from the return address to the patch site is placed just below the return address in the edi slot of the pushad/popad ares. This is safe because the edi register (which is pushed last) is a temporary for the lithium instruction.
As the instanceof stub can call other JavaScript an additional marking for saving all double registers have been added.
Also tweaked the instanceof stub to produce true/false objects instead of 0/1 for the case with deferred code.
Review URL: http://codereview.chromium.org/5990005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6173 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The function HEnvironment::SetExpressionStackAt did not update the
environment's history. This function is used to patch the bailout
environment for count operations and global function calls.
Reorganize class HEnvironment to make it fit V8's style a bit better
and to try to add some sanity to which C++ functions are intended to
be inlined.
Remove the flag --trace-environment which merely duplicated data in
the hydrogen.cfg file except without enough context to be useful.
BUG=1004
Review URL: http://codereview.chromium.org/5992011
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6137 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
1. Separating out the instance-type check from the array-length operation.
2. I also changed the bounds-check on keyed loads to use the length property
for JS arrays (like we do for array stores).
The new pattern should use less registers and allow more checks to be eliminated.
Review URL: http://codereview.chromium.org/5961016
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6125 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This prevents code motion from hoisting map-checks across such stores
which may result in unnecessary deoptimizations.
In the following example program we would move a map-check from the inner loop out before the outer loop which is not desirable:
function f() {
var o = {};
var j = 0;
o.a = 1;
do {
o.b = 6; // Map transition
for (var i=0; i<10; i++) {
o.a = o.b + i;
}
} while(++j < 1) {}
}
for (var i = 0; i < 1000000; i++) f();
Review URL: http://codereview.chromium.org/5991001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6071 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When running with 10 stress runs we get polymorphic stores. This revealed
an incorrect bailout ID for the state following a polymorphic store. The ID
should be the internal ID of the assignment side-effect, not the ID of the
end of the expression.
Review URL: http://codereview.chromium.org/5885003
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6051 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Before, when we deoptimized after a branch we jumped to before the branch
was taken in the unoptimized code with a token value that indicated when
edge to take. There was a lot of machinery to track this value through the
short-circuit logical operations and logical negation, and to handle it
properly at inline function return sites. There was also machinery to
prevent incorrectly seeing this environment with the extra value never
actually materialized in the unoptimized code.
Instead, now we deoptimize directly to one of the targets of the branch.
Much but not yet all of the extra machinery has been removed or simplified.
The cost is that branching control structures (the looping statements, if
statements, conditional expressions, and the short-circuit binary logical
operations) need extra AST IDs to identify the branch targets.
Review URL: http://codereview.chromium.org/5908001
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6049 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The issue is caused when deoptimizing to an internal AST ID in a
postfix increment or decrement operation on variable. This could
happen for a global variable. In that case, the optimized code was
not properly simulating an extra stack slot in the unoptimized code to
hold the original value.
Review URL: http://codereview.chromium.org/5871002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6025 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This fixes V8 issue 989.
Before, assignments used the AST ID of the assignment expression to
mark the side effect of the store, which became a target for
deoptimization bailout for code after the assignment. In effect
contexts this environment included the value of the assignment, which
was unexpected by the unoptimized code.
Now we introduce a new assignment ID for AST node types that include
an assignment (Assignment, CountOperation, and ForInStatement) and use
it for the side effect of the store.
Review URL: http://codereview.chromium.org/5682010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5990 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
When falling off the end of a function inlined in a test context, we cannot
constant fold the test of undefined away. The graph builder assumes that
control flow always reaches both branches of a test.
Instead, explicitly test and branch on "undefined". Introduce a pair of
empty blocks to hold the necessary LeaveInlined instructions.
Review URL: http://codereview.chromium.org/5566005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5927 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
