The idea is that you can create a VarDecl with an unknown type, or a
FunctionDecl with an unknown return type, and it will still be valid to
access that object as long as you explicitly cast it at every use. I'm
still going back and forth about how I want to test this effectively, but
I wanted to go ahead and provide a skeletal implementation for the LLDB
folks' benefit and because it also improves some diagnostic goodness for
placeholder expressions.
llvm-svn: 129065
developers can see if their driver changed any cl::Option's. The
current implementation isn't perfect but handles most kinds of
options. This is nice to have when decomposing the stages of
compilation and moving between different drivers. It's also a good
sanity check when comparing results produced by different command line
invocations that are expected to produce the comparable results.
Note: This is not an attempt to prolong the life of cl::Option. On the
contrary, it's a placeholder for a feature that must exist when
cl::Option is replaced by a more appropriate framework. A new
framework needs: a central option registry, dynamic name lookup,
non-global containers of option values (e.g. per-module,
per-function), *and* the ability to print options values and their defaults at
any point during compilation.
llvm-svn: 128911
the array alignment to the array access.
- This is more or less the best we can do without having alignment present in
the type system, but is a long way from truly matching how GCC handles this.
llvm-svn: 128691
__block object copy/dispose helpers for C++ objects with those for
different variables with completely different semantics simply because
they happen to both be no more aligned than a pointer.
Found by inspection.
Also, internalize most of the helper generation logic within CGBlocks.cpp,
and refactor it to fit my peculiar aesthetic sense.
llvm-svn: 128618
Emit them instead with the linkage of the VTT.
I'm actually really ambivalent about this; it's what GCC does, but outside
of improving code size (if the linkage is coalescing), I'm not sure it's
at all relevant. Construction vtables are naturally referenced only by the
VTT, which is itself only referenced by complete-object constructors and
destructors; giving the construction vtables possibly-external linkage is
important if you have an optimization that drills through the VTT to a
reference to a particular construction vtable which it cannot just emit
itself.
llvm-svn: 128374
platform implies default visibility. To achieve these, refactor our
lookup of explicit visibility so that we search for both an explicit
VisibilityAttr and an appropriate AvailabilityAttr, favoring the
VisibilityAttr if it is present.
llvm-svn: 128336
- Moved the CGObjCRuntime functions out of CGObjCMac.cpp into CGObjCRuntime.cpp
- Added generic functions in CGObjCRuntime for emitting @try and @synchronize
blocks, usable by any runtime that uses DWARF exceptions.
- Made the GNU runtimes use these functions.
It should now be possible to replace the equivalent functions in
CGObjCNonFragileABIMac with simple calls to these two functions, providing the
runtime functions as arguments. I'll post a diff to the list for review before
making any changes to the Mac runtime stuff.
llvm-svn: 128274
accessed via the indirect pointer, they don't need to be pointers to pointers).
Finished moving the message lookup code into separate subclasses for each
runtime. Also performed a few smallish related tidies.
We're now bitcasting the result of the message lookup functions, rather than
casting the lookup functions themselves, so the messages.m test needed updating
to reflect this.
llvm-svn: 128180
that I hadn't used C++ for several years before writing most of this code).
Still lots more to do. This set of changes includes:
- Remove the distinction between typed and untyped selectors. More accurately
reflect what the runtime does, by using typed selectors everywhere, with an
empty type field if the types are unknown. Now we just store a small list of
types for each selector (in theory, this should always be exactly one, but
this constraint was not enforced back in 1986 when it should have been).
- Add some consistency to how runtime functions are created. These are all
generated via the LazyRuntimeFunction class (which might be useful outside
CGObjCGNU - feel free to move it into a header if it is). This function
stores the types of a function, looks it up the first time it's used, and
caches the result. This means that we're now not wasting time constructing
the llvm::FunctionType every time some of the functions are looked up, but
also not inserting references to runtime functions into the module if they're
not actually used.
- Started separating out the fragile and non-fragile ABI behaviours into two
subclasses of CGObjCGNU: CGObjCGCC for the legacy GCC runtime ABI and
CGObjCGNUstep for the new GNUstep ABI. Not all of the differences in
behaviour are factored out yet, but they will be in future commits.
- Removed all of the CodeGen:: things: we've been using namespace CodeGen in
this file for ages, so having explicit namespace specifiers is just a bit
confusing.
- Added a few more comments.
- Used llvm::StringRef instead of std::string in a few places.
- Finally got around to storing the module path in the module structure. The
ABI says that the compiler should do this, although it's not used in the
runtime or exposed outside the runtime, so it's pretty useless.
Still to do:
- We currently have two code paths for generating try blocks, one for ObjC and
one for ObjC++. Not only are these substantially similar, they are also very
similar to the CGObjCMac version. These need factoring out into a single
parameterised implementation, either in CGObjCRuntime or CodeGenFunction.
The EmitObjCXXTryStmt() function was added so that the changes to fix a bug
in time for the 2.9 release would be self-contained and reduce the chances of
breaking anything else, but these should be done properly as soon as
possible.
- Split up some large functions (e.g. GenerateClass()) into smaller functions
for generating the various data structures.
- The method lookup code into the two subclasses, removing the conditionals in
the message send functions.
- Add doxygen comments on the remaining undocumented functions.
- We seem to be generating global pointer variables for selectors, then storing
a pointer to the selector, then generating a load of this pointer (and then a
load of the real selector later) every time a static selector is used. I can
only assume I was asleep or drunk when I did this - we should just be
referencing the selectors directly in the selector array.
llvm-svn: 128152
which versions of an OS provide a certain facility. For example,
void foo()
__attribute__((availability(macosx,introduced=10.2,deprecated=10.4,obsoleted=10.6)));
says that the function "foo" was introduced in 10.2, deprecated in
10.4, and completely obsoleted in 10.6. This attribute ties in with
the deployment targets (e.g., -mmacosx-version-min=10.1 specifies that
we want to deploy back to Mac OS X 10.1). There are several concrete
behaviors that this attribute enables, as illustrated with the
function foo() above:
- If we choose a deployment target >= Mac OS X 10.4, uses of "foo"
will result in a deprecation warning, as if we had placed
attribute((deprecated)) on it (but with a better diagnostic)
- If we choose a deployment target >= Mac OS X 10.6, uses of "foo"
will result in an "unavailable" warning (in C)/error (in C++), as
if we had placed attribute((unavailable)) on it
- If we choose a deployment target prior to 10.2, foo() is
weak-imported (if it is a kind of entity that can be weak
imported), as if we had placed the weak_import attribute on it.
Naturally, there can be multiple availability attributes on a
declaration, for different platforms; only the current platform
matters when checking availability attributes.
The only platforms this attribute currently works for are "ios" and
"macosx", since we already have -mxxxx-version-min flags for them and we
have experience there with macro tricks translating down to the
deprecated/unavailable/weak_import attributes. The end goal is to open
this up to other platforms, and even extension to other "platforms"
that are really libraries (say, through a #pragma clang
define_system), but that hasn't yet been designed and we may want to
shake out more issues with this narrower problem first.
Addresses <rdar://problem/6690412>.
As a drive-by bug-fix, if an entity is both deprecated and
unavailable, we only emit the "unavailable" diagnostic.
llvm-svn: 128127
line options, instead of leveraging the blanket -mllvm option.
- This allows using the frontend itself without requiring the backend have
those options available (i.e., if the target wasn't built).
llvm-svn: 128087
add support for the OpenCL __private, __local, __constant and
__global address spaces, as well as the __read_only, _read_write and
__write_only image access specifiers. Patch originally by ARM;
language-specific address space support by myself.
llvm-svn: 127915
Issue this as an IR-gen error; it's not really worthwhile doing this
"right", i.e. in Sema, because IR gen knows a lot of tricks beyond
what the constant evaluator knows.
llvm-svn: 127854
replace some uses of FieldOffsetInBytes. The remaining uses of
FieldOffsetInBytes will be replaced once NextFieldOffsetInBytes is converted
to CharUnits. No change in functionality intended.
llvm-svn: 127641
Change the interface to expose the new information and deal with the enormous fallout.
Introduce the new ExceptionSpecificationType value EST_DynamicNone to more easily deal with empty throw specifications.
Update the tests for noexcept and fix the various bugs uncovered, such as lack of tentative parsing support.
llvm-svn: 127537
simplify the logic of initializing function parameters so that we don't need
both a variable declaration and a type in FunctionArgList. This also means
that we need to propagate the CGFunctionInfo down in a lot of places rather
than recalculating it from the FAL. There's more we can do to eliminate
redundancy here, and I've left FIXMEs behind to do it.
llvm-svn: 127314
clobber with the 'y' constraint. Otherwise, we get the wrong return type and an
assert, because it created a '<1 x i64>' vector type instead of the x86_mmx
type.
llvm-svn: 127185
21 int main() {
22 A a;
For example, here user would expect to stop at line 22, even if A's constructor leads to a call through CXXDefaultArgExpr.
This fixes ostream-defined.exp regression from gdb testsuite.
llvm-svn: 127164
allocation and therefore requires a null-check. We were doing that, but
we weren't treating the new-initializer as being conditionally executed,
which means it was possible to get ill-formed IR as in PR9298.
llvm-svn: 127147
= bar() + ... + bar() + ...
clang keeps track of column numbers, so we could put location entries for all subexpressions but that will significantly bloat debug info in general, but a location for call expression is helpful here.
llvm-svn: 127018
The prototype for objc_msgSend() is technically variadic -
`id objc_msgSend(id, SEL, ...)`.
But all method calls should use a prototype that matches the method,
not the prototype for objc_msgSend itself().
// rdar://9048030
llvm-svn: 126754
The prototype for objc_msgSend() is technically variadic -
`id objc_msgSend(id, SEL, ...)`.
But all method calls should use a prototype that matches the method,
not the prototype for objc_msgSend itself().
// rdar://9048030
llvm-svn: 126678
C++ exceptions, even when exceptions have been turned off using -fno-exceptions.
Make the -fobjc-exceptions flag do the same thing, but for Objective-C exceptions.
C++ and Objective-C exceptions can also be disabled using -fno-cxx-excptions and
-fno-objc-exceptions.
llvm-svn: 126630
live case of a switch statement when switching on a constant. This is terribly
limited, but enough to handle the trivial example included. Before we would
emit:
define void @test1(i32 %i) nounwind {
entry:
%i.addr = alloca i32, align 4
store i32 %i, i32* %i.addr, align 4
switch i32 1, label %sw.epilog [
i32 1, label %sw.bb
]
sw.bb: ; preds = %entry
%tmp = load i32* %i.addr, align 4
%inc = add nsw i32 %tmp, 1
store i32 %inc, i32* %i.addr, align 4
br label %sw.epilog
sw.epilog: ; preds = %sw.bb, %entry
switch i32 0, label %sw.epilog3 [
i32 1, label %sw.bb1
]
sw.bb1: ; preds = %sw.epilog
%tmp2 = load i32* %i.addr, align 4
%add = add nsw i32 %tmp2, 2
store i32 %add, i32* %i.addr, align 4
br label %sw.epilog3
sw.epilog3: ; preds = %sw.bb1, %sw.epilog
ret void
}
now we emit:
define void @test1(i32 %i) nounwind {
entry:
%i.addr = alloca i32, align 4
store i32 %i, i32* %i.addr, align 4
%tmp = load i32* %i.addr, align 4
%inc = add nsw i32 %tmp, 1
store i32 %inc, i32* %i.addr, align 4
ret void
}
This improves -O0 compile time (less IR to generate and shove through the code
generator) and the clever linux kernel people found a way to fail to build if we
don't do this optimization. This step isn't enough to handle the kernel case
though.
llvm-svn: 126597
necessarily enclose the innermost normal cleanup depth, because
the top of the jump scope stack might be an EH cleanup or EH scope.
Fixes PR9303.
llvm-svn: 126472
invocation function into the debug info. Rather than faking up a class,
which is tricky because of the custom layout we do, we just emit a struct
directly from the layout information we've already got.
Also, don't emit an unnecessarily parameter alloca for this "variable".
llvm-svn: 126255
without defining them. This should be an error, but I'm paranoid about
"uses" that end up not actually requiring a definition. I'll revisit later.
Also, teach IR generation to not set internal linkage on variable
declarations, just for safety's sake. Doing so produces an invalid module
if the variable is not ultimately defined.
Also, fix several places in the test suite where we were using internal
functions without definitions.
llvm-svn: 126016
_Block_object_* flags; it's just BLOCK_HAS_COPY_DISPOSE or not.
Also, we don't need to chase forwarding pointers prior to calling
_Block_object_dispose; _Block_object_dispose in fact already does
this.
rdar://problem/9006315
llvm-svn: 125823
This removes the final dependency edge from any lib outside of CodeGen
to core. As a result we can, and do, trim the dependency on core
from libclang, PrintFunctionNames, the unit tests and c-index-test.
While at it, review and trim other unneeded dependencies.
llvm-svn: 125820
bugs from other clients that don't expect to see a LabelDecl in a DeclStmt,
but if so they should be easy to fix.
This implements most of PR3429 and rdar://8287027
llvm-svn: 125817
class and to bind the shared value using OpaqueValueExpr. This fixes an
unnoticed problem with deserialization of these expressions where the
deserialized form would lose the vital pointer-equality trait; or rather,
it fixes it because this patch also does the right thing for deserializing
OVEs.
Change OVEs to not be a "temporary object" in the sense that copy elision is
permitted.
This new representation is not totally unawkward to work with, but I think
that's really part and parcel with the semantics we're modelling here. In
particular, it's much easier to fix things like the copy elision bug and to
make the CFG look right.
I've tried to update the analyzer to deal with this in at least some
obvious cases, and I think we get a much better CFG out, but the printing
of OpaqueValueExprs probably needs some work.
llvm-svn: 125744
LabelDecl and LabelStmt. There is a 1-1 correspondence between the
two, but this simplifies a bunch of code by itself. This is because
labels are the only place where we previously had references to random
other statements, causing grief for AST serialization and other stuff.
This does cause one regression (attr(unused) doesn't silence unused
label warnings) which I'll address next.
This does fix some minor bugs:
1. "The only valid attribute " diagnostic was capitalized.
2. Various diagnostics printed as ''labelname'' instead of 'labelname'
3. This reduces duplication of label checking between functions and blocks.
Review appreciated, particularly for the cindex and template bits.
llvm-svn: 125733
Nobody ever gave me a clear reason for why we were doing this, and
now it's apparently causing serious problems, so if *not* having this
causes problems, we get to solve them the right way this time.
llvm-svn: 125627
- Have CGM precompute a number of commonly-used types
- Have CGF copy that during initialization instead of recomputing them
- Use TBAA info when initializing a parameter variable
- Refactor the scalar ++/-- code
llvm-svn: 125562
a zero constant for a complete class. rdar://problem/8424975
To make this happen, track the field indexes for virtual bases
in the complete object. I'm curious whether we might be better
off making CGRecordLayoutBuilder *much* more reliant on
ASTRecordLayout; we're currently duplicating an awful lot of the ABI
layout logic.
llvm-svn: 125555
there were only three virtual methods of any significance.
The primary way to grab child iterators now is with
Stmt::child_range children();
Stmt::const_child_range children() const;
where a child_range is just a std::pair of iterators suitable for
being llvm::tie'd to some locals. I've left the old child_begin()
and child_end() accessors in place, but it's probably a substantial
penalty to grab the iterators individually now, since the
switch-based dispatch is kindof inherently slower than vtable
dispatch. Grabbing them together is probably a slight win over the
status quo, although of course we could've achieved that with vtables, too.
I also reclassified SwitchCase (correctly) as an abstract Stmt
class, which (as the first such class that wasn't an Expr subclass)
required some fiddling in a few places.
There are somewhat gross metaprogramming hooks in place to ensure
that new statements/expressions continue to implement
getSourceRange() and children(). I had to work around a recent clang
bug; dgregor actually fixed it already, but I didn't want to
introduce a selfhosting dependency on ToT.
llvm-svn: 125183
and we later find the definition, make sure that we add the definition
(not the declaration) to the list of deferred definitions to
emit. Fixes PR8864.
Thanks to Nick Lewycky for testing this patch out
llvm-svn: 125157
- BlockDeclRefExprs always store VarDecls
- BDREs no longer store copy expressions
- BlockDecls now store a list of captured variables, information about
how they're captured, and a copy expression if necessary
With that in hand, change IR generation to use the captures data in
blocks instead of walking the block independently.
Additionally, optimize block layout by emitting fields in descending
alignment order, with a heuristic for filling in words when alignment
of the end of the block header is insufficient for the most aligned
field.
llvm-svn: 125005
right for anonymous struct/union members led to me discovering some
seemingly broken code in that area of Sema, which I fixed, partly by
changing the representation of member pointer constants so that
IndirectFieldDecls aren't expanded. This led to assorted cleanups with
member pointers in CodeGen, and while I was doing that I saw some random
other things to clean up.
llvm-svn: 124785
is not defined in the current translation unit. Doing so lead to compile errors
such as PR9114.
Instead, when CodeGen is building the vtable, don't try to emit a definition
for functions that aren't marked used in the current translation unit.
Fixes PR9114.
llvm-svn: 124768
* llvm-link would complains about mismatched visibility
* If we produce a relocation with an available_externally, it is good to know that
it is hidden.
llvm-svn: 124633
current translation unit as available_externally.
This helps devirtualize the second example in PR3100, comment 18:
struct S { S() {}; virtual void xyzzy(); };
inline void foo(S *s) { s->xyzzy(); }
void bar() { S s; foo(&s); }
This involved four major changes:
1. In DefineUsedVTables, always mark virtual member functions as referenced for
non-template classes and class template specializations.
2. In CodeGenVTables::ShouldEmitVTableInThisTU return true if optimizations are
enabled, even if the key function is not implemented in this translation
unit. We don't ever do this for code compiled with -fapple-kext, because we
don't ever want to devirtualize virtual member function calls in that case.
3. Give the correct linkage for vtables where the key function is not defined.
4. Update the linkage for RTTI structures when necessary.
llvm-svn: 124565
fixing a crash which probably nobody was ever going to see. In doing so,
fix a horrendous number of problems with the conditional-cleanups code.
Also, make conditional cleanups re-use the cleanup's activation variable,
which avoids some unfortunate repetitiveness.
llvm-svn: 124481
non-class prvalues actually require the realization of a
temporary. For everything else, we already have an lvalue (or class
prvalue) in the subexpression.
Note: we're missing some move elision in this case. I'll tackle that next.
llvm-svn: 124453
deallocation function has a two-argument form. Store the result of this
check in new[] and delete[] nodes.
Fixes rdar://problem/8913519
llvm-svn: 124373
I'm separately committing this because it incidentally changes some
block orderings and minor IR issues, like using a phi instead of
an unnecessary alloca.
llvm-svn: 124277
clang's -Wuninitialized-experimental warning.
While these don't look like real bugs, clang's
-Wuninitialized-experimental analysis is stricter
than GCC's, and these fixes have the benefit
of being general nice cleanups.
llvm-svn: 124072
For example:
class A{
public:
A& operator=(const A& that) {
if (this != &that) {
this->A::~A();
this->A::A(that); // <=== explicit constructor call.
}
return *this;
}
};
More work will be needed to support an explicit call to a template constructor.
llvm-svn: 123735
replace all uses of the entry with the predecessor. There are no cleanups
relying on this right now, but if we ever want a cleanup with a phi inside
it, this will be important.
llvm-svn: 123438
process, perform a number of refactorings:
- Move MiscNameMangler member functions to MangleContext
- Remove GlobalDecl dependency from MangleContext
- Make MangleContext abstract and move Itanium/Microsoft functionality
to their own classes/files
- Implement ASTContext::createMangleContext and have CodeGen use it
No (intended) functionality change.
llvm-svn: 123386
delete the block we began emitting into if it had no predecessors. We never
want to do this, because there are several valid cases during statement
emission where an existing block has no known predecessors but will acquire
some later. The case in my test case doesn't inherently fall into this
category, because we could safely emit the case-range code before the statement
body, but there are examples with labels that can't be fallen into
that would also demonstrate this bug.
rdar://problem/8837067
llvm-svn: 123303
In particular, the iteration variable (if present) should be created and
destroyed in a narrow span around the loop body, and the body should
be emitted in a cleanup scope in case it's not a compound statement.
Otherwise, rename a few variables and use phis instead of temporary
variables for the index and buffer count.
llvm-svn: 122988
The initial TreeTransform is a cop-out, but it's more-or-less equivalent
to what we were doing before, or rather what we're doing now and might
eventually stop doing in favor of using this type.
I am simultaneously intrigued by the possibilities of rebuilding a
dependent Attri
llvm-svn: 122942
expansions with something that is easier to use correctly: a new
template argment kind, rather than a bit on an existing kind. Update
all of the switch statements that deal with template arguments, fixing
a few latent bugs in the process. I"m happy with this representation,
now.
And, oh look! Template instantiation and deduction work for template
template argument pack expansions.
llvm-svn: 122896
the declaration-specifiers and on the declarator itself are moved
to the appropriate declarator chunk. This permits a greatly
simplified model for how to apply these attributes, as well as
allowing a much more efficient query for the GC attribute.
Now all qualifier queries follow the same basic strategy of
"local qualifiers, local qualifiers on the canonical type,
then look through arrays". This can be easily optimized by
changing the canonical qualified-array-type representation.
Do not process type attributes as decl attributes on declarations
with declarators.
When computing the type of a block, synthesize a prototype
function declarator chunk if the decl-spec type was not a
function. This simplifies the logic for building block signatures.
Change the logic which inserts an objc_read_weak on a block
literal to only fire if the block has a __weak __block variable,
rather than if the return type of the block is __weak qualified,
which is not actually a sensible thing to ask.
llvm-svn: 122871
template argument (described by an expression, of course). For
example:
template<int...> struct int_tuple { };
template<int ...Values>
struct square {
typedef int_tuple<(Values*Values)...> type;
};
It also lays the foundation for pack expansions in an initializer-list.
llvm-svn: 122751
in asm statements:
register int foo asm("rdi");
asm("..." : ... "r" (foo) ...
We also only accept these variables if the constraint in the asm statement is "r".
This fixes most of PR3933.
llvm-svn: 122643
16-bits in size. Implement this by splitting WChar into two enums, like we have
for char. This fixes a miscompmilation of XULRunner, PR8856.
llvm-svn: 122558
pack expansions, e.g. given
template<typename... Types> struct tuple;
template<typename... Types>
struct tuple_of_refs {
typedef tuple<Types&...> types;
};
the type of the "types" typedef is a PackExpansionType whose pattern
is Types&.
This commit introduces support for creating pack expansions for
template type arguments, as above, but not for any other kind of pack
expansion, nor for any form of instantiation.
llvm-svn: 122223
Also tweak the VCVT_F32_F16 entry in arm_neon.td to be more consistent with
the other floating-point conversion builtins. Radar 8068427.
llvm-svn: 121916
within the class. Teach IR gen to look for function definitions in record
lexical contexts when deciding whether to emit a function whose address
was taken. Fixes PR8789.
llvm-svn: 121833
class to be passed around. The line between argument and return types and
everything else is kindof vague, but I think it's justifiable.
llvm-svn: 121752
space better. Remove this reference. To make that work, change some APIs
(most importantly, getDesugaredType()) to take an ASTContext& if they
need to return a QualType. Simultaneously, diminish the need to return a
QualType by introducing some useful APIs on SplitQualType, which is
just a std::pair<const Type *, Qualifiers>.
llvm-svn: 121478
zextOrTrunc(), and APSInt methods extend(), extOrTrunc() and new method
trunc(), to be const and to return a new value instead of modifying the
object in place.
llvm-svn: 121121
the LHS, or else the pointer might be invalid. This is kindof dumb, but
go ahead and make sure we're doing that for l-value scalar assignment,
which fixes a miscompile of obj-c++.dg/block-seq.mm.
Leave a FIXME for how to solve this problem for agg __blocks.
llvm-svn: 120992
Fix a bug in the emission of complex compound assignment l-values.
Introduce a method to emit an expression whose value isn't relevant.
Make that method evaluate its operand as an l-value if it is one.
Fixes our volatile compliance in C++.
llvm-svn: 120931
struct X {
X() : au_i1(123) {}
union {
int au_i1;
float au_f1;
};
};
clang will now deal with au_i1 explicitly as an IndirectFieldDecl.
llvm-svn: 120900
not actually frequently used, because ImpCastExprToType only creates a node
if the types differ. So explicitly create an ICE in the lvalue-to-rvalue
conversion code in DefaultFunctionArrayLvalueConversion() as well as several
other new places, and consistently deal with the consequences throughout the
compiler.
In addition, introduce a new cast kind for loading an ObjCProperty l-value,
and make sure we emit those nodes whenever an ObjCProperty l-value appears
that's not on the LHS of an assignment operator.
This breaks a couple of rewriter tests, which I've x-failed until future
development occurs on the rewriter.
Ted Kremenek kindly contributed the analyzer workarounds in this patch.
llvm-svn: 120890
Also, move the l-value emission code into CGObjC.cpp and teach it, for
completeness, to store away self for a super send.
Also, inline the super cases for property gets and sets and make them
use the correct result type for implicit getter/setter calls.
llvm-svn: 120887
when an initializer is variable (I handled the constant case in a previous
patch). This has three pieces:
1. Enhance AggValueSlot to have a 'isZeroed' bit to tell CGExprAgg that
the memory being stored into has previously been memset to zero.
2. Teach CGExprAgg to not emit stores of zero to isZeroed memory.
3. Teach CodeGenFunction::EmitAggExpr to scan initializers to determine
whether they are profitable to emit a memset + inividual stores vs
stores for everything.
The heuristic used is that a global has to be more than 16 bytes and
has to be 3/4 zero to be candidate for this xform. The two testcases
are illustrative of the scenarios this catches. We now codegen test9 into:
call void @llvm.memset.p0i8.i64(i8* %0, i8 0, i64 400, i32 4, i1 false)
%.array = getelementptr inbounds [100 x i32]* %Arr, i32 0, i32 0
%tmp = load i32* %X.addr, align 4
store i32 %tmp, i32* %.array
and test10 into:
call void @llvm.memset.p0i8.i64(i8* %0, i8 0, i64 392, i32 8, i1 false)
%tmp = getelementptr inbounds %struct.b* %S, i32 0, i32 0
%tmp1 = getelementptr inbounds %struct.a* %tmp, i32 0, i32 0
%tmp2 = load i32* %X.addr, align 4
store i32 %tmp2, i32* %tmp1, align 4
%tmp5 = getelementptr inbounds %struct.b* %S, i32 0, i32 3
%tmp10 = getelementptr inbounds %struct.a* %tmp5, i32 0, i32 4
%tmp11 = load i32* %X.addr, align 4
store i32 %tmp11, i32* %tmp10, align 4
Previously we produced 99 stores of zero for test9 and also tons for test10.
This xforms should substantially speed up -O0 builds when it kicks in as well
as reducing code size and optimizer heartburn on insane cases. This resolves
PR279.
llvm-svn: 120692
a global is larger than 32 bytes and has fewer than 6 non-zero values in the
initializer. Previously we'd turn something like this:
char test8(int X) {
char str[10000] = "abc";
into a 10K global variable which we then memcpy'd from. Now we generate:
%str = alloca [10000 x i8], align 16
%tmp = getelementptr inbounds [10000 x i8]* %str, i64 0, i64 0
call void @llvm.memset.p0i8.i64(i8* %tmp, i8 0, i64 10000, i32 16, i1 false)
store i8 97, i8* %tmp, align 16
%0 = getelementptr [10000 x i8]* %str, i64 0, i64 1
store i8 98, i8* %0, align 1
%1 = getelementptr [10000 x i8]* %str, i64 0, i64 2
store i8 99, i8* %1, align 2
Which is much smaller in space and also likely faster.
This is part of PR279
llvm-svn: 120645
-Move the stuff of Diagnostic related to creating/querying diagnostic IDs into a new DiagnosticIDs class.
-DiagnosticIDs can be shared among multiple Diagnostics for multiple translation units.
-The rest of the state in Diagnostic object is considered related and tied to one translation unit.
-Have Diagnostic point to the SourceManager that is related with. Diagnostic can now accept just a
SourceLocation instead of a FullSourceLoc.
-Reflect the changes to various interfaces.
llvm-svn: 119730
store it on the expression node. Also store an "object kind",
which distinguishes ordinary "addressed" l-values (like
variable references and pointer dereferences) and bitfield,
@property, and vector-component l-values.
Currently we're not using these for much, but I aim to switch
pretty much everything calculating l-valueness over to them.
For now they shouldn't necessarily be trusted.
llvm-svn: 119685
of all the lines of the inline asm. With the refactoring and enhancement
of the backend, we can now reports errors on the correct source line when
an asm contains multiple lines of text. For something like this:
void foo() {
asm("push %rax\n"
".code32\n");
}
we used to get this: (note that the line 4 in t.c isn't helpful)
t.c:4:7: error: warning: ignoring directive for now
asm("push %rax\n"
^
<inline asm>:2:1: note: instantiated into assembly here
.code32
^
now we get:
t.c:5:8: error: warning: ignoring directive for now
".code32\n"
^
<inline asm>:2:1: note: instantiated into assembly here
.code32
^
Note that we're pointing to line 5 properly now. This implements
rdar://7839391 - inline asm errors should point to the right line in the asm
and makes the error message in PR8595 much less confusing.
llvm-svn: 119489
assignment to volatiles in C. This in effect reverts some of mjs's
work in and around r72572. Basically, the C++ standard is quite
clear, except that it lies about volatile behavior approximating
C's, whereas the C standard is almost actively misleading.
llvm-svn: 119344
one of the special Neon types. We'll check for invalid Neon vectors when
they are created, so there's no point in handling them when mangling.
llvm-svn: 119299
Return the result of a complex assignment with the original values,
not by performing a load from the l-value; this is the correct
semantics in C, although not in C++.
llvm-svn: 119037
implicit conversions; the last batch was specific to promotions.
I think this is the full set we need. I do think dividing the cast
kinds into floating and integral is probably a good idea.
Annotate a *lot* more C casts with useful cast kinds.
llvm-svn: 119036
NEON vector types need to be mangled in a special way to comply with ARM's ABI,
similar to some of the AltiVec-specific vector types. This patch is mostly
just renaming a bunch of "AltiVecSpecific" things, since they will no longer
be specific to AltiVec. Besides that, it just adds the new "NeonVector" enum.
llvm-svn: 118724
mangler. Now member functions and pointers thereof have their calling
convention mangled as __thiscall if they have the default CC (even though,
they technically still have the __cdecl CC).
llvm-svn: 118598
there's no return adjustment from the overridden to the overrider doesn't
mean there isn't a return adjustment from the overrider to the final
overrider. This matters if we're emitting a virtual this-adjustment thunk
because the overrider virtually inherits from the class providing the
nearest overridden method. Do the appropriate return adjustment in this case.
Fixes PR7611.
llvm-svn: 118466
abstractions (e.g., TemplateArgumentListBuilder) that were designed to
support variadic templates. Only a few remnants of variadic templates
remain, in the parser (parsing template type parameter packs), AST
(template type parameter pack bits and TemplateArgument::Pack), and
Sema; these are expected to be used in a future implementation of
variadic templates.
But don't get too excited about that happening now.
llvm-svn: 118385
data members by delaying the emission of the initializer until after
linkage and visibility have been set on the global. Also, don't
emit a guard unless the variable actually ends up with vague linkage,
and don't use thread-safe statics in any case.
llvm-svn: 118336
independently of whether they're definitions, then teach IR generation to
ignore non-explicit visibility when emitting declarations. Use this to
make sure that RTTI, vtables, and VTTs get the right visibility.
More of rdar://problem/8613093
llvm-svn: 117781
in asm's. PR 8501, 8602988.
I don't like including Type.h where it is; the idea was
to get references to X86_MMXTy out of the common code.
Maybe there's a better way?
llvm-svn: 117736
for namespace-scope variable declarations.
Apply visibility in IR gen to variables that are merely declared
and never defined. We were previously emitting these with default
visibility unless they were declared with private_extern.
Ignore global visibility settings when computing visibility for
a declaration's context, and key several conditions on whether a
visibility attribute exists anywhere in the hierarchy as opposed
to whether it exists at the current level.
llvm-svn: 117729
and never defined. We were previously emitting these with default
visibility unless they were declared with private_extern.
Ignore global visibility settings when computing visibility for
a declaration's context, and key several conditions on whether a
visibility attribute exists anywhere in the hierarchy as opposed
to whether it exists at the current level.
llvm-svn: 117644
in the scope checker. With that done, turn an indirect goto into a
protected scope into a hard error; otherwise IR generation has to start
worrying about declarations not dominating their scopes, as exemplified
in PR8473.
If this really affects anyone, I can probably adjust this to only hard-error
on possible indirect gotos into VLA scopes rather than arbitrary scopes.
But we'll see how people cope with the aggressive change on the marginal
feature.
llvm-svn: 117539
more closely parallel the computation of linkage. This gets us to a state
much closer to what gcc emits, modulo bugs, which will undoubtedly arise in
abundance.
llvm-svn: 117147
function parameters weren't converted to use the correct type (x86_mmx). Add a
check, similar to the one in llvm-gcc, to see if we need the x86_mmx type for
that function parameter. If so, it coerces the type to be that.
llvm-svn: 116684
objc_exception_rethrow, so we don't...", since something is actually trying to
call this with the wrong signature (!). Unfortunately I don't understand the new
EH infrastructure well enough to fix it immediately.
llvm-svn: 116660
__builtin_ia32_vec_init_v8qi
__builtin_ia32_vec_init_v4hi
__builtin_ia32_vec_init_v2si
They are lowered to bitcasts. (These are all ready tested by the gcc testsuite.)
<rdar://problem/8529957>
llvm-svn: 116147
for the same destination, then we must potentially rewrite the initial branch
of every fixup. Without this patch, a short-circuit check meant to prevent
a switch case from being redundantly added was preventing later fixups from
being processed. Fixes PR8175 (again).
llvm-svn: 115586
both @catches and a @finally, because the second call to @objc_exception_try_enter
will clobber the exception slot. Fixes rdar://problem/8440970.
llvm-svn: 115575
force it to be a constant instead of emitting with EmitScalarExpr. In
-ftrapv mode, they are not the same.
This fixes rdar://8478728 + PR8221
llvm-svn: 115388
In this experimental mode try avoiding debug info emission for classes as much as possible. The goal is to reduce size of produced debuginfo without reducing quality of debug info in general. This is a work in progress.
llvm-svn: 115188
one of them) was causing a series of failures:
http://google1.osuosl.org:8011/builders/clang-x86_64-darwin10-selfhost/builds/4518
svn merge -c -114929 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114929 into '.':
U include/clang/Sema/Sema.h
U include/clang/AST/DeclCXX.h
U lib/Sema/SemaDeclCXX.cpp
U lib/Sema/SemaTemplateInstantiateDecl.cpp
U lib/Sema/SemaDecl.cpp
U lib/Sema/SemaTemplateInstantiate.cpp
U lib/AST/DeclCXX.cpp
svn merge -c -114925 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114925 into '.':
G include/clang/AST/DeclCXX.h
G lib/Sema/SemaDeclCXX.cpp
G lib/AST/DeclCXX.cpp
svn merge -c -114924 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114924 into '.':
G include/clang/AST/DeclCXX.h
G lib/Sema/SemaDeclCXX.cpp
G lib/Sema/SemaDecl.cpp
G lib/AST/DeclCXX.cpp
U lib/AST/ASTContext.cpp
svn merge -c -114921 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114921 into '.':
G include/clang/AST/DeclCXX.h
G lib/Sema/SemaDeclCXX.cpp
G lib/Sema/SemaDecl.cpp
G lib/AST/DeclCXX.cpp
llvm-svn: 114933
already be determined by isCopyAssignmentOperator(), and was set too
late in the process for all clients to see the appropriate
value. Cleanup only; no functionality change.
llvm-svn: 114916
- Therefore, we can lower out the NEON wrapper structs and pass the vectors
directly. This makes a huge difference in the cleanliness of the IR after
optimization.
- I will trust, but verify, via future ABITest testing (for APCS-GNU, at
least).
llvm-svn: 114618
the bases are completely initialized. This won't work --- base
initializer expressions can rely on the vtables having been set up.
Check for uses of 'this' in the initializers and force a vtable
initialization if found.
This might not be good enough; we might need to extend this to handle
the possibility of arbitrary code finding an external reference to this
(not yet completely-constructed!) object and accessing through it,
in which case we'll probably find ourselves doing a lot more unnecessary
stores.
llvm-svn: 114153
the cleanup might not be dominated by the allocation code.
In this case, we have to store aside all the delete arguments
in case we need them later. There's room for optimization here
in cases where we end up not actually needing the cleanup in
different branches (or being able to pop it after the
initialization code).
Also make sure we only call this operator delete along the path
where we actually allocated something.
Fixes rdar://problem/8439196.
llvm-svn: 114145
with a non-default-stack-ABI-alignment (of 16).
- This fixes the ABI convenient, but breaks codegen since we now have
underaligned arguments. Marginal improvement overall though, and will be
fixed in next commit.
llvm-svn: 114113
slot. The easiest way to do that was to bundle up the information
we care about for aggregate slots into a new structure which demands
that its creators at least consider the question.
I could probably be convinced that the ObjC 'needs GC' bit should
be rolled into this structure.
Implement generalized copy elision. The main obstacle here is that
IR-generation must be much more careful about making sure that exactly
llvm-svn: 113962
information when imported variable is used
more than once. Originally though to be a bug in importing
block varibles. Fixes radar 8417746.
llvm-svn: 113675
block-literal initializer expression causes IRgen to crash.
This patch fixes by saving it in StaticLocalDecl map
already used for such purposes. (radar 8390455).
llvm-svn: 113307
using the same methods as used for normal structures.
- This fixes problems with reading past the end of the structure and with
handling straddled bit-field access.
llvm-svn: 112914
complains when the element type of a C++ "delete" expression is
different from what we would expect from the pointer type. When
deleting a bool*, we end up with an i1 on one side (where we compute
the LLVM type from the Clang bool type) and i8 on the other (where we
grab the LLVM type from the LLVM pointer type). I've weakened the
assertion appropriately, and the Boost Parallel Graph Library now
passes its regression tests.
llvm-svn: 112821
constructing an LLVM PointerType directly from the "bool"'s LLVM type
(i1), which resulted in unfortunate pointer type i1*. The fix is to
build the LLVM PointerType from the corresponding Clang PointerType,
so that we get i8* in the case of a bool.
John, please review. I also left a FIXME there because we seem to be
dropping "volatile", which would be rather unfortunate.
llvm-svn: 112819
implement ARM array cookies. Also fix a few unfortunate bugs:
- throwing dtors in deletes prevented the allocation from being deleted
- adding the cookie to the new[] size was not being considered for
overflow (and, more seriously, was screwing up the earlier checks)
- deleting an array via a pointer to array of class type was not
causing any destructors to be run and was passing the unadjusted
pointer to the deallocator
- lots of address-space problems, in case anyone wants to support
free store in a variant address space :)
llvm-svn: 112814
with zext/sext operations, instead of to llvm intrinsics. I have a plan to
avoid the clang builtins for these, but it is going to take a little longer
and I want to get the NEON intrinsics updated before the 2.8 release.
llvm-svn: 112764
(and thus protocol_begin(), protocol_end()) now only contains the list of protocols that were directly referenced in
an @interface declaration. 'all_referenced_protocol_[begin,end]()' now returns the set of protocols that were referenced
in both the @interface and class extensions. The latter is needed for semantic analysis/codegen, while the former is
needed to maintain the lexical information of the original source.
Fixes <rdar://problem/8380046>.
llvm-svn: 112691
caused by my ABI work. Passing:
struct outer {
int x;
struct epsilon_matcher {} e;
int f;
};
as {i32,i32} isn't safe, because the offset of the second element
needs to be at 8 when it is interpreted as a memory value.
llvm-svn: 112686
The extra data stored on user-defined literal Tokens is stored in extra
allocated memory, which is managed by the PreprocessorLexer because there isn't
a better place to put it that makes sure it gets deallocated, but only after
it's used up. My testing has shown no significant slowdown as a result, but
independent testing would be appreciated.
llvm-svn: 112458
with zext/sext operations, instead of to llvm intrinsics. (We can also
get rid of the clang builtins and handle these entirely in the arm_neon.h
header if there is a way to express vector sext/zext in C.)
llvm-svn: 112413
but not in C++, so don't emit aggregate loads of volatile references
in null context in C++. Happens to have been caught by an assertion.
We do not get the scalar case right. Volatiles are really broken.
llvm-svn: 112019
That revision started classifying truly empty structs like "Y" and "X"
as being NoClass/NoClass and turning them into 'ignore'. The call code
turns around and allocates space for the ignored argument with
GetUndefRValue. The bug is that GetUndefRValue would return the address
as undef, instead of returning an object with a defined address but
undefined contents.
llvm-svn: 111794
class; they should just be completely opaque throughout IR gen now,
although I haven't really audited that.
Fix a bug apparently inherited from gcc-4.2 where we failed to null-check
member data pointers when performing derived-to-base or base-to-derived
conversions on them.
llvm-svn: 111789
Make CGT defer to the ABI on all member pointer types.
This requires giving CGT a handle to the ABI.
It's way easier to make that work if we avoid lazily creating the ABI.
Make it so.
llvm-svn: 111786
implicitly-defined default constructor, zero-initialize the memory
before calling the default constructor. Previously, we would only
zero-initialize in the case of a trivial default constructor.
Also, simplify the hideous logic that determines when we have a
trivial default constructor and, therefore, don't need to emit any
call at all.
llvm-svn: 111779
pointers. I find the resulting code to be substantially cleaner, and it
makes it very easy to use the same APIs for data member pointers (which I have
conscientiously avoided here), and it avoids a plethora of potential
inefficiencies due to excessive memory copying, but we'll have to see if it
actually works.
llvm-svn: 111776
the ABI code. Implement correct semantics for these on ARM.
I believe this completes the implementation of member function pointers
on ARM.
I think I'm going to switch member function pointers over to be
non-aggregates while I have all this in mind.
llvm-svn: 111774
duplication between the constant and non-constant paths in all of this.
Implement ARM ABI semantics for member pointer constants and conversion.
llvm-svn: 111772
update callers as best I can.
- This is a work in progress, our alignment handling is very horrible / sketchy -- I am just aiming for monotonic improvement.
- Serious review appreciated.
llvm-svn: 111707
not part of the IR, are not uniqued, and may be safely RAUW'd.
This replaces a variety of alternate mechanisms for achieving
the same effect.
llvm-svn: 111682
trivial default constructors. We're weren't zero-initializing them,
which manifested as <rdar://problem/8320532> (a regression in the GCC
test suite) and is likely to have caused significant other breakage.
llvm-svn: 111650
active C++ ABI as a raw string, we store it as an enum. This should improve
performance somewhat.
And yes, this time, I started from a clean build directory, and
all the tests passed. :)
llvm-svn: 111507
Now all classes derived from Attr are generated from TableGen.
Additionally, Attr* is no longer its own linked list; SmallVectors or
Attr* are used. The accompanying LLVM commit contains the updates to
TableGen necessary for this.
Some other notes about newly-generated attribute classes:
- The constructor arguments are a SourceLocation and a Context&,
followed by the attributes arguments in the order that they were
defined in Attr.td
- Every argument in Attr.td has an appropriate accessor named getFoo,
and there are sometimes a few extra ones (such as to get the length
of a variadic argument).
Additionally, specific_attr_iterator has been introduced, which will
iterate over an AttrVec, but only over attributes of a certain type. It
can be accessed through either Decl::specific_attr_begin/end or
the global functions of the same name.
llvm-svn: 111455
mangleCallExpression. Also, operator names with unknown arity should
be mangled as binary operators; this is actually covered by an oddly-
positioned sentence in the ABI document. Fixes PR7891.
llvm-svn: 111395
a -cc1 option. The Darwin linker complains about mixed visibility when linking
gcc-built objects with clang-built objects, and the optimization isn't really
that valuable. Platforms with less ornery linkers can feel free to enable this.
llvm-svn: 110979
instead of _Unwind_Resume. With SJLJ exceptions, this is spelled
"_Unwind_SjLj_Resume_or_Rethrow", not "_Unwind_SjLj_Resume", which has
significantly different semantics.
We should actually never be generating a call to _Unwind_SjLj_Resume directly;
even if we were generating true cleanups (which we aren't because of the
horrible hack), we should be calling __cxa_end_cleanup() on ARM. I
haven't implemented this because there's little point as long as the HH is
present.
I believe this fixes <rdar://problem/8281377>.
llvm-svn: 110851
where we weren't accounting for the possibility that a @finally block might
have internal cleanups and therefore might write to the cleanup destination slot.
Fixes <rdar://problem/8293901>.
llvm-svn: 110760
an lvalue of another, compatible Objective-C object type (e.g., a
subclass). Introduce a new initialization sequence step kind to
describe this binding, along with a new cast kind. Fixes PR7741.
llvm-svn: 110513
to avoid the awesome-but-wrong-in-this-case assertion in the canon EAC.
Fixes PR7834.
Also fix a subtle address-space bug in the memset path.
llvm-svn: 110511
This takes some trickery since CastExpr has subclasses (and indeed,
is abstract).
Also, smoosh the CastKind into the bitfield from Expr.
Drops two words of storage from Expr in the common case of expressions
which don't need inheritance paths. Avoids a separate allocation and
another word of overhead in cases needing inheritance paths. Also has
the advantage of not leaking memory, since destructors for AST nodes are
never run.
llvm-svn: 110507
__overflow_handler entrypoint that David Chisnall made up.
Calling __overflow_handler is not part of the contract of
-ftrapv provided by GCC, and should never have been checked
in in the first place.
According to:
http://permalink.gmane.org/gmane.comp.compilers.clang.devel/8699
David is using this for some of arbitrary precision integer stuff
or something, which is not an appropriate thing to implement on
this.
llvm-svn: 110490
do the right thing with mixed-visibility symbols, so disable the visibility
optimization where that's possible, i.e. with template classes (since it's
possible that an arbitrary template might be subject to an explicit
instantiation elsewhere). 447.dealII actually does this.
I've put the code under an option that's currently not hooked up to anything.
llvm-svn: 110374
functions with in-line definitions, since such thunks will be emitted at any
use of the function.
Completes the feature work for rdar://problem/7523229.
llvm-svn: 110285
Apply hidden visibility to most RTTI; libstdc++ does not rely on exact
pointer equality for the type info (just the type info names). Apply
the same optimization to RTTI that we do to vtables.
Fixes PR5962.
llvm-svn: 110192
ObjC exceptions:
- don't enter a try for the catch blocks unless there's a finally
- put the setjmp buffer in the locals set for liveness reasons
- dump the sync object into an alloca in the locals set for liveness reasons
Some of this can go away if the backend starts to properly calculate liveness
in the presence of setjmp (which would also be a *much* stabler solution).
llvm-svn: 110188
the magic of inline assembly. Essentially we use read and write hazards
on the set of local variables to force flushing locals to memory
immediately before any protected calls and to inhibit optimizing locals
across the setjmp->catch edge. Fixes rdar://problem/8160285
llvm-svn: 109960
The X86-64 ABI code didn't handle the case when a struct
would get classified and turn up as "NoClass INTEGER" for
example. This is perfectly possible when the first slot
is all padding (e.g. due to empty base classes). In this
situation, the first 8-byte doesn't take a register at all,
only the second 8-byte does.
This fixes this by enhancing the x86-64 abi stuff to allow
and handle this case, reverts the broken fix for PR5831,
and enhances the target independent stuff to be able to
handle an argument value in registers being accessed at an
offset from the memory value.
This is the last x86-64 calling convention related miscompile
that I'm aware of.
llvm-svn: 109848
sections on", this change uncovered a possible linker bug which resulted in the
wrong messages getting dispatched. Backing this out while we investigate...
llvm-svn: 109817
<2 x float> instead of double. This works but can't be turned
on until I teach codegen to pass <2 x float> as one XMM register
instead of two.
llvm-svn: 109790
DeclIsRequiredFunctionOrFileScopedVar.
This is essentially a CodeGen predicate that is also needed by the PCH mechanism to determine whether a decl
needs to be deserialized during PCH loading for codegen purposes.
Since this logic is shared by CodeGen and the PCH mechanism, move it to the ASTContext,
thus CodeGenModule's GetLinkageForFunction/GetLinkageForVariable and the GVALinkage enum is moved out of CodeGen.
This fixes current (and avoids future) codegen-from-PCH bugs.
llvm-svn: 109784
return where the struct has a base but no fields. This
was because the x86-64 abi logic was checking the wrong
predicate in one place.
This was introduced in r91874, which was a fix for PR5831,
which lacked a CHECK line, so I verified and added it.
llvm-svn: 109759
have a "coerce to" type which often matches the default lowering of Clang
type to LLVM IR type, but the coerce case can be handled by making them
not be the same.
This simplifies things and fixes issues where X86-64 abi lowering would
return coerce after making preferred types exactly match up. This caused
us to compile:
typedef float v4f32 __attribute__((__vector_size__(16)));
v4f32 foo(v4f32 X) {
return X+X;
}
into this code at -O0:
define <4 x float> @foo(<4 x float> %X.coerce) nounwind {
entry:
%retval = alloca <4 x float>, align 16 ; <<4 x float>*> [#uses=2]
%coerce = alloca <4 x float>, align 16 ; <<4 x float>*> [#uses=2]
%X.addr = alloca <4 x float>, align 16 ; <<4 x float>*> [#uses=3]
store <4 x float> %X.coerce, <4 x float>* %coerce
%X = load <4 x float>* %coerce ; <<4 x float>> [#uses=1]
store <4 x float> %X, <4 x float>* %X.addr
%tmp = load <4 x float>* %X.addr ; <<4 x float>> [#uses=1]
%tmp1 = load <4 x float>* %X.addr ; <<4 x float>> [#uses=1]
%add = fadd <4 x float> %tmp, %tmp1 ; <<4 x float>> [#uses=1]
store <4 x float> %add, <4 x float>* %retval
%0 = load <4 x float>* %retval ; <<4 x float>> [#uses=1]
ret <4 x float> %0
}
Now we get:
define <4 x float> @foo(<4 x float> %X) nounwind {
entry:
%X.addr = alloca <4 x float>, align 16 ; <<4 x float>*> [#uses=3]
store <4 x float> %X, <4 x float>* %X.addr
%tmp = load <4 x float>* %X.addr ; <<4 x float>> [#uses=1]
%tmp1 = load <4 x float>* %X.addr ; <<4 x float>> [#uses=1]
%add = fadd <4 x float> %tmp, %tmp1 ; <<4 x float>> [#uses=1]
ret <4 x float> %add
}
This implements rdar://8248065
llvm-svn: 109733
Before we'd compile the example into something like:
%coerce.dive2 = getelementptr %struct.v4f32wrapper* %retval, i32 0, i32 0 ; <<4 x float>*> [#uses=1]
%1 = bitcast <4 x float>* %coerce.dive2 to <2 x double>* ; <<2 x double>*> [#uses=1]
%2 = load <2 x double>* %1, align 1 ; <<2 x double>> [#uses=1]
ret <2 x double> %2
Now we produce:
%coerce.dive2 = getelementptr %struct.v4f32wrapper* %retval, i32 0, i32 0 ; <<4 x float>*> [#uses=1]
%0 = load <4 x float>* %coerce.dive2, align 1 ; <<4 x float>> [#uses=1]
ret <4 x float> %0
llvm-svn: 109732
possible. This improves the example to pass <4 x float> instead of
<2 x double> but we still get awful code, and still don't get the
return value right.
llvm-svn: 109700
enclosing normal cleanup, not the top of the EH stack. I'm *really*
surprised this hasn't been causing more problems.
Fixes rdar://problem/8231514.
llvm-svn: 109569
DeclIsRequiredFunctionOrFileScopedVar.
This function is part of the public CodeGen interface since it's essentially a CodeGen predicate that is also
needed by the PCH mechanism to determine whether a decl needs to be deserialized during PCH loading for codegen purposes.
This fixes current (and avoids future) codegen-from-PCH bugs.
llvm-svn: 109546
use of property-dot syntax using 'super' as receiver
is 'void'. This fixes a bug in generating correct
API for setter call. Fixes radar 8203426.
llvm-svn: 109297
them as such. Type::is(Signed|Unsigned|)IntegerType() now return false
for vector types, and new functions
has(Signed|Unsigned|)IntegerRepresentation() cover integer types and
vector-of-integer types. This fixes a bunch of latent bugs.
Patch from Anton Yartsev!
llvm-svn: 109229
initializer of (). Make sure to use a simple memset() when we can, or
fall back to generating a loop when a simple memset will not
suffice. Fixes <rdar://problem/8212208>, a regression due to my work
in r107857.
llvm-svn: 108977
which generates more efficient and more obviously conformant
code. We now test for overflow of the multiply then force
the result to -1 if so. On X86, this generates nice code
like this:
__Z4testl: ## @_Z4testl
## BB#0: ## %entry
subl $12, %esp
movl $4, %eax
mull 16(%esp)
testl %edx, %edx
movl $-1, %ecx
cmovel %eax, %ecx
movl %ecx, (%esp)
call __Znam
addl $12, %esp
ret
llvm-svn: 108927
causing clang to compile this code into something that correctly throws a
length error, fixing a potential integer overflow security attack:
void *test(long N) {
return new int[N];
}
int main() {
test(1L << 62);
}
We do this even when exceptions are disabled, because it is better for the
code to abort than for the attack to succeed.
This is heavily based on a patch that Fariborz wrote.
llvm-svn: 108915
their call expressions synthetically have the "deduced" types based on their
first argument. We only insert conversions in the AST for arguments whose
values require conversion to match the value type expected. This keeps PR7600
closed by maintaining the return type, but avoids assertions due to unexpected
implicit casts making the type unsigned (test case added from Daniel).
The magic is moved into the codegen for the atomic builtin which inserts the
casts as needed at the IR level to raise the type to an integer suitable for
the LLVM intrinsic. This shouldn't cause any real change in functionality, but
now we can make the builtin be more truly polymorphic.
llvm-svn: 108638
- This issue here is that /usr/include/Blocks.h wants to define some of the
block runtime globals as weak, depending on the target. This doesn't work in
Clang because we aren't using the AST decl for these globals.
- The fix is a pretty gross hack which just watches all the decls for the
specific blocks globals we need to know about; if we see one we use it,
otherwise we use the hand coded type.
In time, I would like to clean this up by changing IRgen to ask Sema/AST for
the decl, which would then be lazily loaded from the builtin table if
necessary. This could be used in a whole host of places in IRgen and would
get rid of a lot of grotty hand coding of LLVM IR; however, we need some
extra Sema support for this as well as support for builtin global variables.
llvm-svn: 108482
that involve binding a reference to a pure rvalue temporary (e.g., not
a class temporary), by creating a new temporary and copying the result
there. Fixes PR6024.
llvm-svn: 108431
whether to use objc_msgSend_fpret; the choice is target dependent, not Obj-C ABI
dependent.
- <rdar://problem/8139758> arm objc _objc_msgSend_fpret bug
llvm-svn: 108379
- TSTs whose template is a template template parameter already work
- we don't provide an imaginary type, so we can't mangle one
- we don't need a generic FIXME for vendor type qualifiers
llvm-svn: 108317
reinterpret_casts (possibly indirectly via C-style/functional casts)
on values, e.g.,
int i;
reinterpret_cast<short&>(i);
The IR generated for this is essentially the same as for
*reinterpret_cast<short*>(&i).
Fixes PR6437, PR7593, and PR7344.
llvm-svn: 108294
mostly in avoiding unnecessary work at compile time but also in producing more
sensible block orderings.
Move the destructor cleanups for local variables over to use lazy cleanups.
Eventually all cleanups will do this; for now we have some awkward code
duplication.
Tell IR generation just to never produce landing pads in -fno-exceptions.
This is a much more comprehensive solution to a problem which previously was
half-solved by checks in most cleanup-generation spots.
llvm-svn: 108270
(e.g., a call, cast, etc.), immediately adjust the expression's type
to strip cv-qualifiers off of all non-class types (in C++) or all
types (in C). This effectively extends my previous fix for PR7463,
which was restricted to calls, to other kinds of expressions within
similar characteristics. I've audited every use of
getNonReferenceType() in the code base, switching to the newly-renamed
getNonLValueExprType() where necessary.
Big thanks to Eli for pointing out just how incomplete my original fix
for PR7463 actually was. We've been handling cv-qualifiers on rvalues
wrong for a very, very long time. Fixes PR7463.
llvm-svn: 108253
at -O0. The only change from the previous patch is that we don't try
to generate virtual method thunks for an available_externally
function.
llvm-svn: 108230
-O0, since we won't be using the definitions for anything anyway. For
lib/System/Path.o when built in Debug+Asserts mode, this leads to a 4%
improvement in compile time (and suppresses 440 function bodies).
<rdar://problem/7987644>
llvm-svn: 108156
was not producing a memcpy with the right address
spaces because of two places in it doing casts of
the arguments to i8, one of which that didn't
preserve the address space.
There is also an optimizer bug here.
llvm-svn: 107842
CXXConstructExpr/CXXTemporaryObjectExpr/CXXNewExpr as
appropriate. Fixes PR7556, and provides a slide codegen improvement
when copy-initializing a POD class type from a value-initialized
temporary. Previously, we weren't eliding the copy.
llvm-svn: 107827
emit metadata associating allocas and global values with a Decl*. This feature
is controlled by an option that (intentionally) cannot be enabled on the command
line.
To use this feature, simply set
CodeGenOptions.EmitDeclMetadata = true;
and then interpret the completely underspecified metadata. :)
llvm-svn: 107739
self-host. Hopefully these results hold up on different platforms.
I tried to keep the GNU ObjC runtime happy, but it's hard for me to test.
Reimplement how clang generates IR for exceptions. Instead of creating new
invoke destinations which sequentially chain to the previous destination,
push a more semantic representation of *why* we need the cleanup/catch/filter
behavior, then collect that information into a single landing pad upon request.
Also reorganizes how normal cleanups (i.e. cleanups triggered by non-exceptional
control flow) are generated, since it's actually fairly closely tied in with
the former. Remove the need to track which cleanup scope a block is associated
with.
Document a lot of previously poorly-understood (by me, at least) behavior.
The new framework implements the Horrible Hack (tm), which requires every
landing pad to have a catch-all so that inlining will work. Clang no longer
requires the Horrible Hack just to make exceptions flow correctly within
a function, however. The HH is an unfortunate requirement of LLVM's EH IR.
llvm-svn: 107631
alloca for an argument. Make sure the argument gets the proper
decl alignment, which may be different than the type alignment.
This fixes PR7567
llvm-svn: 107627
ObjC pointers were easy enough (as far as the ABI is concerned, they're
just pointers to structs), but I had to invent a new mangling for block
pointers. This is particularly worrying with the Microsoft ABI, because
it is a vendor-specific ABI; extending it could come back to bite us
later when MS extends it on their own (and you know they will).
llvm-svn: 107572
Only actual functions get mangled correctly; I don't know how to fix it for
function pointers yet. Thanks to John McCall for the hint.
Also, mangle anonymous tag types. I don't have a suitable testcase yet; I have
a feeling that that's going to need support for static locals, and I haven't
figured out exactly how MSVC's scheme for mangling those works.
llvm-svn: 107561
store make sure to move the debug metadata from the store (which is actual
'return' statement location) to the return instruction (which otherwise would
have the function end location as its debug info).
- Tested by gdb test suite.
llvm-svn: 107322
r107173, "fix PR7519: after thrashing around and remembering how all this stuff"
r107216, "fix PR7523, which was caused by the ABI code calling ConvertType instead"
This includes a fix to make ConvertTypeForMem handle the "recursive" case, and call
it as such when lowering function types which have an indirect result.
llvm-svn: 107310
doesn't mangle array parameters right), but I think that should be fixed
in Sema (Doug, John, what do you think?).
Also, stub out the remaining mangleType() routines.
llvm-svn: 107264
complex values either. Previously we did this properly for regular assignment,
but not for compound assignment.
- Also, tidy up assignment code a bit to look more like the scalar path.
llvm-svn: 107217
would trigger an extra method call).
- While in the area, I also changed Clang to not emit an unnecessary load from
'x' in cases like 'y = (x = 1)'.
llvm-svn: 107210
works, the fix is quite simple: just make sure to call ConvertTypeRecursive
when the function type being lowered is in the midst of ConvertType.
llvm-svn: 107173
This is somewhat annoying to do this at this level, but it avoids
having ABIInfo know depend on CodeGenTypes for a hint.
Nothing is using this yet, so no functionality change.
llvm-svn: 107111
have CGF create and make accessible standard int32,int64 and
intptr types. This fixes a ton of 80 column violations
introduced by LLVMContextification and cleans up stuff a lot.
llvm-svn: 106977
load/store nonsense in the epilog. For example, for:
int foo(int X) {
int A[100];
return A[X];
}
we used to generate:
%arrayidx = getelementptr inbounds [100 x i32]* %A, i32 0, i64 %idxprom ; <i32*> [#uses=1]
%tmp1 = load i32* %arrayidx ; <i32> [#uses=1]
store i32 %tmp1, i32* %retval
%0 = load i32* %retval ; <i32> [#uses=1]
ret i32 %0
}
which codegen'd to this code:
_foo: ## @foo
## BB#0: ## %entry
subq $408, %rsp ## imm = 0x198
movl %edi, 400(%rsp)
movl 400(%rsp), %edi
movslq %edi, %rax
movl (%rsp,%rax,4), %edi
movl %edi, 404(%rsp)
movl 404(%rsp), %eax
addq $408, %rsp ## imm = 0x198
ret
Now we generate:
%arrayidx = getelementptr inbounds [100 x i32]* %A, i32 0, i64 %idxprom ; <i32*> [#uses=1]
%tmp1 = load i32* %arrayidx ; <i32> [#uses=1]
ret i32 %tmp1
}
and:
_foo: ## @foo
## BB#0: ## %entry
subq $408, %rsp ## imm = 0x198
movl %edi, 404(%rsp)
movl 404(%rsp), %edi
movslq %edi, %rax
movl (%rsp,%rax,4), %eax
addq $408, %rsp ## imm = 0x198
ret
This actually does matter, cutting out 2000 lines of IR from CGStmt.ll
for example.
Another interesting effect is that altivec.h functions which are dead
now get dce'd by the inliner. Hence all the changes to
builtins-ppc-altivec.c to ensure the calls aren't dead.
llvm-svn: 106970
As part of this, pull together trapv handling into the same enum.
This also add support for NSW multiplies.
This also makes PCH disagreement on overflow behavior silent, since it
really doesn't matter except for warnings and codegen (no macros get
defined etc).
llvm-svn: 106956
types, updating callers of both isFloatingType() and
isRealFloatingType() accordingly. Caught at least one issue where we
allowed one to declare a vector of vectors (!), along with cleaning up
the standard-conversion logic for C++.
llvm-svn: 106595
provides C "integer type" semantics in C and C++ "integral type"
semantics in C++.
Note that I still need to update isIntegerType (and possibly other
predicates) using the same approach I've taken for
isIntegralType(). The two should have the same meaning, but currently
don't (!).
llvm-svn: 106074
in C++ that involve both integral and enumeration types. Convert all
of the callers to Type::isIntegralType() that are meant to work with
both integral and enumeration types over to
Type::isIntegralOrEnumerationType(), to prepare to eliminate
enumeration types as integral types.
llvm-svn: 106071
objective-c++ class objects which have GC'able objc object
pointers and need to use ObjC's objc_memmove_collectable
API (radar 8070772).
llvm-svn: 106061
Currently, all AST consumers are located in the Frontend library,
meaning that in a shared library configuration, Frontend has a
dependency on Rewrite, Checker and CodeGen. This is suboptimal for
clients which only wish to make use of the frontend. CodeGen in
particular introduces a large number of unwanted dependencies.
This patch breaks the dependency by moving all AST consumers with
dependencies on Rewrite, Checker and/or CodeGen to their respective
libraries. The patch therefore introduces dependencies in the other
direction (i.e. from Rewrite, Checker and CodeGen to Frontend).
After applying this patch, Clang builds correctly using CMake and
shared libraries ("cmake -DBUILD_SHARED_LIBS=ON").
N.B. This patch includes file renames which are indicated in the
patch body.
Changes in this revision of the patch:
- Fixed some copy-paste mistakes in the header files
- Modified certain aspects of the coding to comply with the LLVM
Coding Standards
llvm-svn: 106010
- Mangle qualifiers.
- Start mangling variables' types into the name. A variable declared with a
builtin type should now mangle properly.
llvm-svn: 105931
(but not their types; that's later).
NOTE: Right now, variables in the global namespace don't get mangled, even
though they're supposed to be. This is because the default mangler
implements the shouldMangleDeclName() method that tells clang not to mangle
them. This will be fixed in a later patch.
llvm-svn: 105805
case of an elaborated-type-specifier like 'typename A<T>::foo', and
DependentTemplateSpecializationType represents the case of an
elaborated-type-specifier like 'typename A<T>::template B<T>'. The TypeLoc
representation of a DependentTST conveniently exactly matches that of an
ElaboratedType wrapping a TST.
Kill off the explicit rebuild methods for RebuildInCurrentInstantiation;
the standard implementations work fine because the nested name specifier
is computable in the newly-entered context.
llvm-svn: 105801
The macros required for DeclNodes use have changed to match the use of
StmtNodes. The FooFirst enumerator constants have been named firstFoo
to match usage elsewhere.
llvm-svn: 105165
isn't possible to compute.
This patch is mostly refactoring; the key change is the addition of the code
starting with the comment, "Check whether the function has a computable LLVM
signature." The solution here is essentially the same as the way the
vtable code handles such functions.
llvm-svn: 105151
This class only supports name mangling (which is apparently used during C/ObjC
codegen). For now only the Itanium C++ ABI is supported. Patches to add a
second C++ ABI are forthcoming.
llvm-svn: 104630
variables within blocks. We loosely follow GCC's mangling, but since
these are always internal symbols the names don't really matter. I
intend to revisit block mangling later, because GCC's mangling is
rather verbose. <rdar://problem/8015719>.
llvm-svn: 104610
variables should have that linkage. Otherwise, its static local
variables should have internal linkage. To avoid computing this excessively,
set a function's linkage before we emit code for it.
Previously we were assigning weak linkage to the static variables of
static inline functions in C++, with predictably terrible results. This
fixes that and also gives better linkage than 'weak' when merging is required.
llvm-svn: 104581
This works around a crash where malloc reused the memory of an erased BB for a
new BB leaving old cleanup information pointing at the new block.
llvm-svn: 104472
temporaries. There are actually several interrelated fixes here:
- When converting an object to a base class, it's only an lvalue
cast when the original object was an lvalue and we aren't casting
pointer-to-derived to pointer-to-base. Previously, we were
misclassifying derived-to-base casts of class rvalues as lvalues,
causing various oddities (including problems with reference binding
not extending the lifetimes of some temporaries).
- Teach the code for emitting a reference binding how to look
through no-op casts and parentheses directly, since
Expr::IgnoreParenNoOpCasts is just plain wrong for this. Also, make
sure that we properly look through multiple levels of indirection
from the temporary object, but destroy the actual temporary object;
this fixes the reference-binding issue mentioned above.
- Teach Objective-C message sends to bind the result as a temporary
when needed. This is actually John's change, but it triggered the
reference-binding problem above, so it's included here. Now John
can actually test his return-slot improvements.
llvm-svn: 104434
not make copies non-POD arguments or arguments passed by reference:
just copy the pointers directly. This eliminates another source of the
dreaded memcpy-of-non-PODs. Fixes PR7188.
llvm-svn: 104327
'self' variable arising from uses of the 'super' keyword. Also reorganize
some code so that BlockInfo (now CGBlockInfo) can be opaque outside of
CGBlocks.cpp.
Fixes rdar://problem/8010633.
llvm-svn: 104312
particular issue was the cause of the Boost.Interprocess failures, and
in general will lead to horrendous, hard-to-diagnose miscompiles. The
assertion itself has survives self-host and a full Boost build, so we
are close to eradicating this problem in C++.
Note that the assertion is *not* turned on for Objective-C++, where we
still have problems with introducing memcpy's of non-POD class
types. That part of the assertion will go away as soon as we fix the
known issues in Objective-C++.
llvm-svn: 104227
subobject. Previously, we could only properly bind to a base class
subobject while extending the lifetime of the complete object (of a
derived type); for non-static data member subobjects, we could memcpy
(!) the result and bind to that, which is rather broken.
Now, we pull apart the expression that we're binding to, to figure out
which subobject we're accessing, then construct the temporary object
(adding a destruction if needed) and, finally, dig out the subobject
we actually meant to access.
This fixes yet another instance where we were memcpy'ing rather than
doing the right thing. However, note the FIXME in references.cpp:
there's more work to be done for binding to subobjects, since the AST
is incorrectly modeling some member accesses in base classes as
lvalues when they are really rvalues.
llvm-svn: 104219
class type (that uses a return slot), pass the return slot to the
callee directly rather than allocating new storage and trying to copy
the object. This appears to have been the cause of the remaining two
Boost.Interprocess failures.
llvm-svn: 104215
This is still probably wrong for Objective-C++ and adds a couple of lines in CGException that should probably be in the CGObjCRuntime subclass. The personality function is now only looked up in one place in CGException though, so this should be easier to fix in the future.
llvm-svn: 103938
__cxa_guard_abort along the exceptional edge into (in effect) a nested
"try" that rethrows after aborting. Fixes PR7144 and the remaining
Boost.ProgramOptions failures, along with the regressions that r103880
caused.
The crucial difference between this and r103880 is that we now follow
LLVM's little dance with the llvm.eh.exception and llvm.eh.selector
calls, then use _Unwind_Resume_or_Rethrow to rethrow.
llvm-svn: 103892
__cxa_guard_abort along the exceptional edge into (in effect) a nested
"try" that rethrows after aborting. Fixes PR7144 and the remaining
Boost.ProgramOptions failures.
llvm-svn: 103880
ObjCObjectType, which is basically just a pair of
one of {primitive-id, primitive-Class, user-defined @class}
with
a list of protocols.
An ObjCObjectPointerType is therefore just a pointer which always points to
one of these types (possibly sugared). ObjCInterfaceType is now just a kind
of ObjCObjectType which happens to not carry any protocols.
Alter a rather large number of use sites to use ObjCObjectType instead of
ObjCInterfaceType. Store an ObjCInterfaceType as a pointer on the decl rather
than hashing them in a FoldingSet. Remove some number of methods that are no
longer used, at least after this patch.
By simplifying ObjCObjectPointerType, we are now able to easily remove and apply
pointers to Objective-C types, which is crucial for a certain kind of ObjC++
metaprogramming common in WebKit.
llvm-svn: 103870
return statements. We perform NRVO only when all of the return
statements in the function return the same variable. Fixes some link
failures in Boost.Interprocess (which is relying on NRVO), and
probably improves performance for some C++ applications.
llvm-svn: 103867
return value optimization. Sema marks return statements with their
NRVO candidates (which may or may not end up using the NRVO), then, at
the end of a function body, computes and marks those variables that
can be allocated into the return slot.
I've checked this locally with some debugging statements (not
committed), but there won't be any tests until CodeGen comes along.
llvm-svn: 103865
"return" statement and mark the corresponding CXXConstructExpr as
elidable. Teach CodeGen that eliding a temporary is different from
eliding an object construction.
This is just a baby step toward NRVO.
llvm-svn: 103849
throw, it should use invoke when needed. The fixes the
Boost.Statechrt failures that motivated PR7132, but there are a few
side issues to tackle as well.
llvm-svn: 103803
methods for which the key function is guaranteed to be in another
translation unit. Unfortunately, this guarantee isn't the case when
dealing with shared libraries that fail to export these virtual method
definitions.
I'm reopening PR6747 so we can consider this again at a later point in
time.
llvm-svn: 103741
"used" (e.g., we will refer to the vtable in the generated code) and
when they are defined (i.e., because we've seen the key function
definition). Previously, we were effectively tracking "potential
definitions" rather than uses, so we were a bit too eager about emitting
vtables for classes without key functions.
The new scheme:
- For every use of a vtable, Sema calls MarkVTableUsed() to indicate
the use. For example, this occurs when calling a virtual member
function of the class, defining a constructor of that class type,
dynamic_cast'ing from that type to a derived class, casting
to/through a virtual base class, etc.
- For every definition of a vtable, Sema calls MarkVTableUsed() to
indicate the definition. This happens at the end of the translation
unit for classes whose key function has been defined (so we can
delay computation of the key function; see PR6564), and will also
occur with explicit template instantiation definitions.
- For every vtable defined/used, we mark all of the virtual member
functions of that vtable as defined/used, unless we know that the key
function is in another translation unit. This instantiates virtual
member functions when needed.
- At the end of the translation unit, Sema tells CodeGen (via the
ASTConsumer) which vtables must be defined (CodeGen will define
them) and which may be used (for which CodeGen will define the
vtables lazily).
From a language perspective, both the old and the new schemes are
permissible: we're allowed to instantiate virtual member functions
whenever we want per the standard. However, all other C++ compilers
were more lazy than we were, and our eagerness was both a performance
issue (we instantiated too much) and a portability problem (we broke
Boost test cases, which now pass).
Notes:
(1) There's a ton of churn in the tests, because the order in which
vtables get emitted to IR has changed. I've tried to isolate some of
the larger tests from these issues.
(2) Some diagnostics related to
implicitly-instantiated/implicitly-defined virtual member functions
have moved to the point of first use/definition. It's better this
way.
(3) I could use a review of the places where we MarkVTableUsed, to
see if I missed any place where the language effectively requires a
vtable.
Fixes PR7114 and PR6564.
llvm-svn: 103718
This fixes recent regressions reported by gdb testsuite.
Tighter verification of debug info generated by FE found these regressions.
Refactor code to extract line number and column number from SourceLocation.
llvm-svn: 103678
available_externally linkage, since they may not have been given a
strong definition in another translation unit. Without this patch, the
following test case fails to link with a GCC-compiled libstdc++:
#include <sstream>
int main() { std::basic_stringbuf<char> bs; }
Fixes the last problem with the Boost.IO library.
llvm-svn: 103208
function attributes like byval get applied to the function
definition. This fixes PR7058 and makes i386 llvm/clang bootstrap
pass all the same tests as x86-64 bootstrap for me (the llvmc
tests still fail in both).
llvm-svn: 103131
reference type, make sure that the initializer we build is the
of the appropriate type for the *reference*, not for the thing that it
refers to. Fixes PR7050.
llvm-svn: 103115
destructors, place the __cxa_atexit call after the __cxa_guard_release
call, mimicking GCC/LLVM-GCC behavior. Noticed while debugging
something related.
llvm-svn: 103088
implicitly-generated copy constructor. Previously, Sema would perform
some checking and instantiation to determine which copy constructors,
etc., would be called, then CodeGen would attempt to figure out which
copy constructor to call... but would get it wrong, or poke at an
uninstantiated default argument, or fail in other ways.
The new scheme is similar to what we now do for the implicit
copy-assignment operator, where Sema performs all of the semantic
analysis and builds specific ASTs that look similar to the ASTs we'd
get from explicitly writing the copy constructor, so that CodeGen need
only do a direct translation.
However, it's not quite that simple because one cannot explicit write
elementwise copy-construction of an array. So, I've extended
CXXBaseOrMemberInitializer to contain a list of indexing variables
used to copy-construct the elements. For example, if we have:
struct A { A(const A&); };
struct B {
A array[2][3];
};
then we generate an implicit copy assignment operator for B that looks
something like this:
B::B(const B &other) : array[i0][i1](other.array[i0][i1]) { }
CodeGen will loop over the invented variables i0 and i1 to visit all
elements in the array, so that each element in the destination array
will be copy-constructed from the corresponding element in the source
array. Of course, if we're dealing with arrays of scalars or class
types with trivial copy-assignment operators, we just generate a
memcpy rather than a loop.
Fixes PR6928, PR5989, and PR6887. Boost.Regex now compiles and passes
all of its regression tests.
Conspicuously missing from this patch is handling for the exceptional
case, where we need to destruct those objects that we have
constructed. I'll address that case separately.
llvm-svn: 103079
they're unreachable. This matters because (if they're POD, or if this is C)
the scope containing the variable might be reachable even if the variable
isn't. Fixes PR7044.
llvm-svn: 103052
typedef int functype(int, int);
functype func;
also instantiate the synthesized function parameters for the resulting
function declaration.
With this change, Boost.Wave builds and passes all of its regression
tests.
llvm-svn: 103025
not just the inner expression. This is important if the expression has any
temporaries. Fixes PR 7028.
Basically a symptom of really tragic method names.
llvm-svn: 102998
assignment operators.
Previously, Sema provided type-checking and template instantiation for
copy assignment operators, then CodeGen would synthesize the actual
body of the copy constructor. Unfortunately, the two were not in sync,
and CodeGen might pick a copy-assignment operator that is different
from what Sema chose, leading to strange failures, e.g., link-time
failures when CodeGen called a copy-assignment operator that was not
instantiation, run-time failures when copy-assignment operators were
overloaded for const/non-const references and the wrong one was
picked, and run-time failures when by-value copy-assignment operators
did not have their arguments properly copy-initialized.
This implementation synthesizes the implicitly-defined copy assignment
operator bodies in Sema, so that the resulting ASTs encode exactly
what CodeGen needs to do; there is no longer any special code in
CodeGen to synthesize copy-assignment operators. The synthesis of the
body is relatively simple, and we generate one of three different
kinds of copy statements for each base or member:
- For a class subobject, call the appropriate copy-assignment
operator, after overload resolution has determined what that is.
- For an array of scalar types or an array of class types that have
trivial copy assignment operators, construct a call to
__builtin_memcpy.
- For an array of class types with non-trivial copy assignment
operators, synthesize a (possibly nested!) for loop whose inner
statement calls the copy constructor.
- For a scalar type, use built-in assignment.
This patch fixes at least a few tests cases in Boost.Spirit that were
failing because CodeGen picked the wrong copy-assignment operator
(leading to link-time failures), and I suspect a number of undiagnosed
problems will also go away with this change.
Some of the diagnostics we had previously have gotten worse with this
change, since we're going through generic code for our
type-checking. I will improve this in a subsequent patch.
llvm-svn: 102853
Amadini.
This change introduces a new expression node type, OffsetOfExpr, that
describes __builtin_offsetof. Previously, __builtin_offsetof was
implemented using a unary operator whose subexpression involved
various synthesized array-subscript and member-reference expressions,
which was ugly and made it very hard to instantiate as a
template. OffsetOfExpr represents the AST more faithfully, with proper
type source information and a more compact representation.
OffsetOfExpr also has support for dependent __builtin_offsetof
expressions; it can be value-dependent, but will never be
type-dependent (like sizeof or alignof). This commit introduces
template instantiation for __builtin_offsetof as well.
There are two major caveats to this patch:
1) CodeGen cannot handle the case where __builtin_offsetof is not a
constant expression, so it produces an error. So, to avoid
regressing in C, we retain the old UnaryOperator-based
__builtin_offsetof implementation in C while using the shiny new
OffsetOfExpr implementation in C++. The old implementation can go
away once we have proper CodeGen support for this case, which we
expect won't cause much trouble in C++.
2) __builtin_offsetof doesn't work well with non-POD class types,
particularly when the designated field is found within a base
class. I will address this in a subsequent patch.
Fixes PR5880 and a bunch of assertions when building Boost.Python
tests.
llvm-svn: 102542
Emitted some metadata on message sends to allow a later pass to do some speculative inlining of class methods (GNU runtime). Speculative inlining of instance methods requires type feedback to be useful (work in progress), but for class methods it works quite nicely.
llvm-svn: 102514
T::template apply<U>), handling a few cases where we previously failed
and performing substitutions on such dependent names. Fixes a crash in
Boost.PropertyTree.
llvm-svn: 102490
This works around stack corruption / crashes resulting from PR6944, and also
works around people who expect 'what works on my machine' to work everywhere
(GCC crashes in a number of cases on SPARC that should now work correctly with
clang).
llvm-svn: 102430
of a class template or class template partial specialization. That is to
say, in
template <class T> class A { ... };
or
template <class T> class B<const T*> { ... };
make 'A<T>' and 'B<const T*>' sugar for the corresponding InjectedClassNameType
when written inside the appropriate context. This allows us to track the
current instantiation appropriately even inside AST routines. It also allows
us to compute a DeclContext for a type much more efficiently, at some extra
cost every time we write a template specialization (which can be optimized,
but I've left it simple in this patch).
llvm-svn: 102407
Ken Dyck