an iterator, since the implementation returned an iterator that pointed to a
different node! Renamed this implementation to SlimFind() so that users do not
expect it to return an iterator (it is a more efficient implementation than
returning an iterator if the user just wants to find the value of a key).
Added a FIXME to implement ImmutableMap::find() that returns an iterator.
llvm-svn: 46150
as weak globals rather than commons. While not wrong,
this change tickled a latent bug in Darwin's strip,
so revert it for now as a workaround.
llvm-svn: 46147
as weak globals rather than commons. While not wrong,
this change tickled a latent bug in Darwin's strip,
so revert it for now as a workaround.
llvm-svn: 46144
is because GNU-style Statement-expressions cause the last statement in the
statement-expression to act like an expression.
We now have two notions: block-level statements and block-level expressions.
The former are all Stmt* that appear in the list of statements in CFGBlocks. The
latter is the subset of the former; these block-level statements are used as
subexpressions somewhere in the AST. CFG::isBlockExpr() returns true for the
latter, not the former (previously isBlockExpr() always returned true for
non-Expr Stmt*).
Modified the LiveVariables analysis to also track liveness state for block-level
expressions (using the updated definition of block-level expressions).
Modified the dataflow solver so that when it records values for block-level
statements, it records the dataflow value *before* the transfer function for a
Stmt* is evaluated (not after). This is more in sync in what clients will want.
Modified CFGStmtVisitor to record the current block-level statement.
llvm-svn: 46143
Fixed CellSPU's A-form (local store) address mode, so that all globals,
externals, constant pool and jump table symbols are now wrapped within
a SPUISD::AFormAddr pseudo-instruction. This now identifies all local
store memory addresses, although it requires a bit of legerdemain during
instruction selection to properly select loads to and stores from local
store, properly generating "LQA" instructions.
Also added mul_ops.ll test harness for exercising integer multiplication.
llvm-svn: 46142
1. Legalize now always promotes truncstore of i1 to i8.
2. Remove patterns and gunk related to truncstore i1 from targets.
3. Rename the StoreXAction stuff to TruncStoreAction in TLI.
4. Make the TLI TruncStoreAction table a 2d table to handle from/to conversions.
5. Mark a wide variety of invalid truncstores as such in various targets, e.g.
X86 currently doesn't support truncstore of any of its integer types.
6. Add legalize support for truncstores with invalid value input types.
7. Add a dag combine transform to turn store(truncate) into truncstore when
safe.
The later allows us to compile CodeGen/X86/storetrunc-fp.ll to:
_foo:
fldt 20(%esp)
fldt 4(%esp)
faddp %st(1)
movl 36(%esp), %eax
fstps (%eax)
ret
instead of:
_foo:
subl $4, %esp
fldt 24(%esp)
fldt 8(%esp)
faddp %st(1)
fstps (%esp)
movl 40(%esp), %eax
movss (%esp), %xmm0
movss %xmm0, (%eax)
addl $4, %esp
ret
llvm-svn: 46140
values for the block-level expressions.
Modified 'LiveVariables' to provide the option to clients to record
liveness information for block-level expressions (using the above feature).
Modified 'DeadStores' to conform to the new interface of 'LiveVariables'.
Modified 'GRConstants' to compute liveness information for block-level
expressions.
llvm-svn: 46137
__builtin_overload takes 2 or more arguments:
0) a non-zero constant-expr for the number of arguments the overloaded
functions will take
1) the arguments to pass to the matching overloaded function
2) a list of functions to match.
The return type of __builtin_overload is inferred from the function whose args
match the types of the arguments passed to the builtin. For example:
float a;
float sinf(float);
int sini(int);
float b = __builtin_overload(1, a, sini, sinf);
Says that we are overloading functions that take one argument, and trying to
pass an argument of the same type as 'a'. sini() does not match since it takes
and argument of type int. sinf does match, so at codegen time this will turn
into float b = sinf(a);
llvm-svn: 46132
and not just the key value when comparing trees. To do this we added data_type
and data_type_ref to the ImutContainerInfo trait classes. For values stored in
the tree that do not have separate key and data components, data_type is simply
a typedef of bool, and isDataEqual() always evaluates to true. This allows us to
support both ImmutableSet and ImmutableMap using the same underlying logic.
llvm-svn: 46130
Refactored the use of this method into both the Sema module and Analysis module,
which were using their own static functions that did the same thing.
llvm-svn: 46129
and switch various codegen pieces and the X86 backend over
to using it.
* Add some comments to SelectionDAGNodes.h
* Introduce a second argument to FP_ROUND, which indicates
whether the FP_ROUND changes the value of its input. If
not it is safe to xform things like fp_extend(fp_round(x)) -> x.
llvm-svn: 46125
and the spill is its kill. However, if the local allocator has determined the
register has not been modified (possible when its value was reloaded), it would
not issue a restore. In that case, mark the last use of the virtual register as
kill.
llvm-svn: 46111
It's not safe to use the two value CombineTo variant to combine away a dead load.
e.g.
v1, chain2 = load chain1, loc
v2, chain3 = load chain2, loc
v3 = add v2, c
Now we replace use of v1 with undef, use of chain2 with chain1.
ReplaceAllUsesWith() will iterate through uses of the first load and update operands:
v1, chain2 = load chain1, loc
v2, chain3 = load chain1, loc
v3 = add v2, c
Now the second load is the same as the first load, SelectionDAG cse will ensure
the use of second load is replaced with the first load.
v1, chain2 = load chain1, loc
v3 = add v1, c
Then v1 is replaced with undef and bad things happen.
llvm-svn: 46099
it should work, but I have no machine to test
it on. Committed because it will at least
cause no harm, and maybe someone can test it
for me!
llvm-svn: 46098
handle the case where the number of nodes was 0.
Fixed bug in GREngineImpl where we did not proceed to the next statement
when processing a PostStmt location.
llvm-svn: 46093
Ted Kremenek