This affects both the Itanium and Microsoft C++ ABIs.
This is in anticipation of a change to the Itanium C++ ABI, and should
match GCC's current behavior. The new text will likely be:
"""
Pass an object of class type by value if every copy constructor and
move constructor is deleted or trivial and at least one of them is not
deleted, and the destructor is trivial.
"""
http://sourcerytools.com/pipermail/cxx-abi-dev/2014-May/002728.html
On x86 Windows, we can mostly use the same logic, where we use inalloca
instead of passing by address. However, on Win64, there are register
parameters, and we have to do what MSVC does. MSVC ignores the presence
of non-trivial move constructors and only considers the presence of
non-trivial or deleted copy constructors. If a non-trivial or deleted
copy ctor is present, it passes the argument indirectly.
This change fixes bugs and makes us more ABI compatible with both GCC
and MSVC.
Fixes PR19668.
Reviewers: rsmith
Differential Revision: http://reviews.llvm.org/D3660
llvm-svn: 208786
In the Microsoft C++ ABI, instance methods always return records
indirectly via the second hidden parameter. This was implemented in
X86_32ABIInfo, but not WinX86_64ABIInfo.
Rather than exposing a handful of boolean methods in the CGCXXABI
interface, we can expose a single method that applies C++ ABI return
value classification rules.
llvm-svn: 208733
Also tidy up, simplify, and extend the test coverage to demonstrate the
limitations. This test should now fail if the bugs are fixed (&
hopefully whoever ends up in this situation sees the FIXMEs and realizes
that the test needs to be updated to positively test their change that
has fixed some or all of these issues).
I do wonder whether I could demonstrate breakage without a macro here,
but any way I slice it I can't think of a way to get two calls to the
same function on the same line/column in non-macro C++ - implicit
conversions happen at the same location as an explicit function, but
you'd never get an implicit conversion on the result of an explicit call
to the same implicit conversion operator (since the value is already
converted to the desired result)...
llvm-svn: 208468
Summary:
MSVC always passes 'sret' after 'this', unlike GCC. This required
changing a number of places in Clang that assumed the sret parameter was
always first in LLVM IR.
This fixes win64 MSVC ABI compatibility for methods returning structs.
Reviewers: rsmith, majnemer
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D3618
llvm-svn: 208458
It was set by default on Darwin in r198655. The same usability issues
with DTrace and LLDB apply to FreeBSD, so set it by default there too.
rdar://problem/15758808
http://llvm.org/pr19676
Differential Revision: http://reviews.llvm.org/D3448
llvm-svn: 208310
r154191 switched to atexit() instead of global destructors, so the intent
was probably to check for _GLOBAL__D_a _not_ being in the output. There already
is a line for _ZN3barD1Ev further up, so just remove the CH_ECK line referring
to that.
The only circumstance in which clang emits _GLOBAL__D_a destructor symbols is
for -fapple-kext, and that is tested by test/CodeGenCXX/cxx-apple-kext.cpp.
llvm-svn: 208222
Summary:
Update our mangling to match the discussion on cxx-abi-dev.
This involves using a seq-id instead of an optional number.
Reviewers: rsmith
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D3631
llvm-svn: 208140
This makes it easier to see where a global ctor comes from, and it also makes
ASan's init order analyzer output easier to understand. gcc does this too,
but only in -fPIC mode for some reason. Don't do this for constructors with
explicit init priority.
Also prepend "sub_" before the 'I', that way regular constructors stay
lexicographically after symbols with init priority (because
ord('s') > ord('I')). gold seems to ignore the name of constructor symbols,
and ld only looks at the symbol if it includes an init priority, which this
patch doesn't change.
Before: __GLOBAL_I_a
Now: __GLOBAL_sub_I_myfile.cc
llvm-svn: 208128
Reverting r208106 to reapply r208065 with a fix for the regression. The
issue was that the enum tried to be built even if the declaration hadn't
been constructed for debug info - presenting problems for enum templates
and typedefs of enums with names for linkage purposes.
Original commit message:
This regressed a little further 208055 though it was already a little
broken.
While the requiresCompleteType optimization should be implemented here.
Future (possibly near future) work.
llvm-svn: 208114
This regressed a little further 208055 though it was already a little
broken.
While the requiresCompleteType optimization should be implemented here.
Future (possibly near future) work.
llvm-svn: 208065
Currently, users get error messages about RTTI descriptor mangling with
no useful source location. This addresses that.
Another approach would be to disable C++ exceptions by default in the
driver when using the Microsoft C++ ABI. However, this makes it
impossible to parse system headers that use exception handling
constructs. By delaying the error to IRgen, we can figure out if we
actually need to emit code for this construct. Additionally, users who
are only interested in building refactoring tools on Windows still get a
correct AST without having to add flags. Finally, this is consistent
with what we do for SEH.
llvm-svn: 207999
We would sometimes incorrectly give a thread-wrapper external linkage
instead of internal linkage if we had only CodeGen'd it's declaration,
not it's definition.
This fixes PR19655.
llvm-svn: 207988
Passing objects directly (in registers or memory) creates a second copy
of the object in the callee. The callee always destroys its copy, but
we also have to destroy any temporary created in the caller. In other
words, copy elision of these kinds of objects is impossible.
Objects larger than 8 bytes with non-trivial dtors and trivial copy
ctors are still passed indirectly, and we can still elide copies of
them.
Fixes PR19640.
llvm-svn: 207889
dependent-type-member-pointer.cpp is failing on a win64 bot because
-fms-extensions is not enabled. Use ConvertType rather than relying on
the inheritance attributes. It's less code, but probably slower.
llvm-svn: 207819
The Win64 ABI docs on MSDN say that arguments bigger than 8 bytes are
passed by reference. Prior to this change, we were only applying this
logic to RecordType arguments. This affects both the Itanium and
Microsoft C++ ABIs.
Reviewers: majnemer
Differential Revision: http://reviews.llvm.org/D3587
llvm-svn: 207817
This code is trying to test if the pointer is *not* null. Therefore we
should use 'or' instead of 'and' to combine the results of 'icmp ne'.
This logic is consistent with the general member pointer comparison code
in EmitMemberPointerComparison.
llvm-svn: 207815
To simplify source location offsets, this test uses line directives to
force particular lines of interest to have known line numbers so that
adjustments before/after those points don't require updates to the CHECK
lines.
llvm-svn: 207779
Summary:
Previously, we would generate a single name for all reference
temporaries and allow LLVM to rename them for us. Instead, number the
reference temporaries as we build them in Sema.
Reviewers: rsmith
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D3554
llvm-svn: 207776
We accept 'void *p; p->~void();' for MSVC compatibility since r148682.
However, we were returning ExprError, rather than producing an AST,
despite only diagnosing it with a warning. CodeGen noticed that the
template function specialization had an invalid AST, and therefore
didn't generate code for it. This change makes us produce an AST with a
void pseudo-dtor call.
Part of PR18256.
llvm-svn: 207771
We were destroying them in the callee, and then again in the caller. We
should use an EH-only cleanup and disable it at the point of the call
for win64, even though we don't use inalloca.
llvm-svn: 207733
It is possible that a variable template specialization might not have a
VisibilityAttr attached to it while the template that it specializes
does, in fact, have one.
We should consider the template in such cases.
This fixes PR19597.
llvm-svn: 207498
Summary:
A reference temporary should inherit the linkage of the variable it
initializes. Otherwise, we may hit cases where a reference temporary
wouldn't have the same value in all translation units.
Reviewers: rsmith
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D3515
llvm-svn: 207451
Prior to this patch, CGRecordLower assumed that virtual bases could not
be placed before the nvsize of an object. This isn't true in Itanium
mode, virtual bases are placed at dsize rather than vnsize and in the
case of zero sized non-virtual bases nvsize can be larger than dsize.
This patch fixes CGRecordLowering to avoid an assert and to clip
bitfields properly in this case. A test case is included.
llvm-svn: 207280
after we've already instantiated a definition for the function, pass it to the
ASTConsumer again so that it knows the specialization kind has changed and can
update the function's linkage.
This only matters if we instantiate the definition of the function before we
reach the end of the TU; this can happen in at least three different ways:
C++11 constexpr functions, C++14 deduced return types, and functions
instantiated within modules.
llvm-svn: 207152
Summary:
Correct size_t to be unsigned int and ptrdiff_t to be signed long. The types were the correct size before this change but
the exact type matters for name mangling and exception handling in C++.
Reviewers: atanasyan
Reviewed By: atanasyan
Differential Revision: http://reviews.llvm.org/D3470
llvm-svn: 207093
We were using the same guard variable and failing to initialize the
second global.
Clang is still not MS ABI compatible in this area. Fixing that is
PR16959, which will require LLVM changes to @llvm.global_ctors.
llvm-svn: 207008
We previously treated ARM separately from the generic Itanium ABI for
initializing guard variables. This code duplication led to things like
the ARM path missing the memory barrier for threadsafe handling, and a
highly misleading comment about how we were (mis)using the generic ABI
for ARM64 when really it went through the ARM codepath.
This unifies the two code paths. Functionally, this changes the ARM
and ARM64 codepath to use one byte loads instead of 4 and 8,
respectively, and adds the missing atomic acquire to these loads.
Other architectures are unchanged.
llvm-svn: 206937
specializations collect all arguments and not just the ones from the
class template partial specialization from which this class template
specialization was instantiated. The debug info does not represent the
partial specialization otherwise and so specialized parameters would
go missing.
rdar://problem/16636569.
llvm-svn: 206430
for CXXGlobalInit/Dtor helper functions.
This makes _GLOBAL__I_a regain its DW_AT_high/low_pc in the debug info.
Thanks to echristo for catching this!
llvm-svn: 206088
are not associated with any source lines.
Previously, if the Location of a Decl was empty, EmitFunctionStart would
just keep using CurLoc, which would sometimes be correct (e.g., thunks)
but in other cases would just point to a hilariously random location.
This patch fixes this by completely eliminating all uses of CurLoc from
EmitFunctionStart and rather have clients explicitly pass in a
SourceLocation for the function header and the function body.
rdar://problem/14985269
llvm-svn: 205999
sure that a debugger can find them when stepping through code,
for example from the included testcase:
12 int test_it() {
13 c = 1;
14 d = 2;
-> 15 a = 4;
16 return (c == 1);
17 }
18
(lldb) p a
(int) $0 = 2
(lldb) p c
(int) $1 = 2
(lldb) p d
(int) $2 = 2
and a, c, d are all part of the file static anonymous union:
static union {
int c;
int d;
union {
int a;
};
struct {
int b;
};
};
Fixes PR19221.
llvm-svn: 205952
If we crash, we raise a crash handler dialog, and that's really
annoying. Even though we can't emit correct IR until we have musttail,
don't crash.
llvm-svn: 205948
cxx-abi-dev came up with a way to disambiguate between different
keywords used in elaborated type specifiers.
This resolves certain collisions during mangling.
llvm-svn: 205943
It is very similar to GCC's __PRETTY_FUNCTION__, except it prints the
calling convention.
Reviewers: majnemer
Differential Revision: http://reviews.llvm.org/D3311
llvm-svn: 205780
This can actually be non-zero if you override a function from a virtual
base and you have forced the most_general pointer to member
representation.
llvm-svn: 205727
We already got the type alias correct (though I've included a test case
here) since Clang represents that like any other typedef - but type
alias templates weren't being handled.
llvm-svn: 205691
While the folding set would deduplicate the nodes themselves and LLVM
would handle not emitting the same global twice, it still meant creating
a long/redundant list of global variables.
llvm-svn: 205668
See the comment for CodeGenFunction::tryEmitAsConstant that describes
how in some contexts (lambdas) we must not emit references to the
variable, but instead use the constant directly - because of this we end
up emitting a constant for the variable, as well as emitting the
variable itself.
Should we just skip putting the variable on the stack at all and omit
the debug info for the constant? It's not clear to me - what if the
address of the local is taken?
llvm-svn: 205651
By ignoring this pragma with a warning, we're essentially miscompiling
the user's program. WebKit / Blink use this pragma to disable dynamic
initialization and finalization of some static data, and running the
dtors crashes the program.
Error out for now, so that /fallback compiles the TU correctly with
MSVC. This pragma should be implemented some time this month, and we
can remove this hack.
llvm-svn: 205554
Summary:
MSVC always emits inline functions marked with the extern storage class
specifier. The result is something similar to the opposite of
__attribute__((gnu_inline)).
This extension is also available in C.
This fixes PR19264.
Reviewers: rnk, rsmith
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3207
llvm-svn: 205485
This was committed 4 years ago in 108916 with insufficient testing to
explain why the "getTypeAsWritten" case was appropriate. Experience says
that it isn't - the presence or absence of an explicit instantiation
declaration was causing this code to generate either i<int> or i<int,
int>.
That didn't seem to be a useful distinction, and omitting the template
arguments was destructive to debuggers being able to associate the two
types across translation units or across compilers (GCC, reasonably,
never omitted the arguments).
llvm-svn: 205447
For namespaces, this is consistent with mangling and GCC's debug info
behavior. For structs, GCC uses <anonymous struct> but we prefer
consistency between all anonymous entities but don't want to confuse
them with template arguments, etc, so we'll just go with parens in all
cases.
llvm-svn: 205398
We don't want to encourage the code to emit a lexical block for
a function that needs one in order for the line table to change,
we need to grab the line information from the body of the pattern
that we were instantiated from, this code should do that.
Modify the test case to ensure that we're still looking in the
right place for all of the scopes and also that we haven't
created a lexical block where we didn't need one.
llvm-svn: 205368
The MS ABI forces us into catch-22 when it comes to functions which
return types which are local:
- A function is mangled with it's return type.
- A type is mangled with it's surrounding context.
Avoid this by mangling auto and decltype(autp) directly into the
function's return type. Using this mangling has the double advantage of
being compatible with the C++ standard without crashing the compiler.
N.B. For the curious, the MSVC mangling leads to collisions amongst
template functions and either crashes when faced with local types or is
otherwise incapable of returning them.
llvm-svn: 205282
Clang implements the part of the ARM ABI saying that certain functions
(e.g., constructors and destructors) return "this", but Apple's version of
gcc and llvm-gcc did not. The libstdc++ dylib on iOS 5 was built with
llvm-gcc, which means that clang cannot safely assume that code from the C++
runtime will correctly follow the ABI. It is also possible to run into this
problem when linking with other libraries built with gcc or llvm-gcc. Even
though there is no way to reliably detect that situation, it is most likely
to come up when targeting older versions of iOS. Disabling the optimization
for any code targeting iOS 5 solves the libstdc++ problem and has a reasonably
good chance of fixing the issue for other older libraries as well.
<rdar://problem/16377159>
llvm-svn: 205272
This adds coverage for Unicode code points which are encoded with
non-zero values in the upper half of the wchar_t.
No functionality change.
llvm-svn: 205251
Summary:
The definition of a type later in a translation unit may change it's
type from {}* to (%struct.foo*)*. Earlier function definitions may use
the former while more recent definitions might use the later. This is
fine until they interact with one another (like one calling the other).
In these cases, a bitcast is needed because the inalloca must match the
function call but the store to the lvalue which initializes the argument
slot has to match the rvalue's type.
This technique is along the same lines with what the other,
non-inalloca, codepaths perform.
This fixes PR19287.
Reviewers: rnk
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3224
llvm-svn: 205217
This adds Clang support for the ARM64 backend. There are definitely
still some rough edges, so please bring up any issues you see with
this patch.
As with the LLVM commit though, we think it'll be more useful for
merging with AArch64 from within the tree.
llvm-svn: 205100
This follows the LLVM change to canonicalise the Windows target triple
spellings. Rather than treating each Windows environment as a single entity,
the environments are now modelled properly as an environment. This is a
mechanical change to convert the triple use to reflect that change.
llvm-svn: 204978
instead of rolling an inefficient version of the function. This
changes some order of emission of metadata nodes, fix up those
testcases and make them more flexible to some changes.
llvm-svn: 204874
COFF doesn't have mergeable sections so LLVM/clang's normal tactics for
string deduplication will not have any effect.
To remedy this we place each string inside it's own section and mark
the section as IMAGE_COMDAT_SELECT_ANY. However, we can only do this if the
string has an external name that we can generate from it's contents.
To be compatible with MSVC, we must use their scheme. Otherwise identical
strings in translation units from clang may not be deduplicated with
translation units in MSVC.
This fixes PR18248.
N.B. We will not attempt to do anything with a string literal which is not of
type 'char' or 'wchar_t' because their compiler does not support unicode
string literals as of this date. Further, we avoid doing this if
either -fwritable-strings or -fsanitize=address are present.
This reverts commit r204596.
llvm-svn: 204675
This commit cleans up a few accidents:
- Do not rely on the order in which StringLiteral lays out bytes.
- Use a more efficient mechanism for handling so-called
"special-mappings" when mangling string literals.
- There is no need to allocate a copy of the mangled name.
- Add the test written for r204562.
Thanks to Richard Smith for pointing these out!
llvm-svn: 204586
location that the next call emitLocation() would default to. Otherwise
setLocation() may wrongly believe that the current source file didn't
change, when in fact it did.
llvm-svn: 204517
complete. We hook into this check from a couple of other places (modules,
debug info) so it's not OK to elide the check if the type was already
complete.
llvm-svn: 203978
These tests are logically related, but they're spread about several
different CodeGen directories. Consolidate them in one place to make
them easier to manage.
llvm-svn: 203541
Previously, we would always emit them with internal linkage,
but with hidden visibility when the function was hidden, which
is an illegal combination, which could lead LLVM to actually
emit them as strong hidden symbols with hilarious results.
rdar://16265084
llvm-svn: 203503
These tests were added before we had settled on using a .profdata extension
for the profile data files. Renaming them now for consistency.
llvm-svn: 203166
If a guard variable will be created for an entity at global scope,
then we cannot rely on the scope depth to disambiguate names for us.
Instead, mangle the entire variable into the guard to ensure it's uniqueness.
llvm-svn: 203151
Initializers and finalizers for static data members have the variable's
access-specifier, storage-class, type and CV-qualifiers mangled in.
llvm-svn: 203145
Use a scheme inspired by the Itanium ABI to properly implement the
mangling of lambdas.
N.B. The incredibly astute observer will notice that we do not generate
external names that are identical, or even compatible with, MSVC.
This is fine because they don't generate names that they can use across
translation units. Technically, we can generate any name we'd like so
long as that name wouldn't conflict with any other and would be stable
across translation units.
This fixes PR15512.
llvm-svn: 202962
Summary:
The MSVC ABI appears to mangle the lexical scope into the names of
statics. Specifically, a counter is incremented whenever a scope is
entered where things can be declared in such a way that an ambiguity can
arise. For example, a class scope inside of a class scope doesn't do
anything interesting because the nested class cannot collide with
another nested class.
There are problems with this scheme:
- It is unreliable. The counter is only incremented when a previously
never encountered scope is entered. There are cases where this will
cause ambiguity amongst declarations that have the same name where one
was introduced in a deep scope while the other was introduced right
after in the previous lexical scope.
- It is wasteful. Statements like: {{{{{{{ static int foo = a; }}}}}}}
will make the mangling of "foo" larger than it need be because the
scope counter has been incremented many times.
Because of these problems, and practical implementation concerns. We
choose not to implement this scheme if the local static or local type
isn't visible. The mangling of these declarations will look very
similar but the numbering will make far more sense, this scheme is
lifted from the Itanium ABI implementation.
Reviewers: rsmith, doug.gregor, rnk, eli.friedman, cdavis5x
Reviewed By: rnk
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2953
llvm-svn: 202951
We wouldn't recognize variable templates as being templates leading us
to leave the template arguments off of the mangled name. This would
allow two unrelated templates to map to the same mangled name.
N.B. While MSVC doesn't support variable templates as of this date,
this mangling is the most likely thing they will choose to use. Their
demangler can successfully demangle our manglings with the template
arguments shown.
llvm-svn: 202789
We should only have this optimization fire when the explicit
instantiation definition would cause at least one member function to be
emitted, thus ensuring that even a compiler not performing this
optimization would still emit the full type information elsewhere.
But we should also pessimize output still by always emitting the
definition when the explicit instantiation definition appears so that at
some point in the future we can depend on that information even when no
code had to be emitted in that TU. (this shouldn't happen very often,
since people mostly use explicit spec decl/defs to reduce code size -
but perhaps one day they could use it to explicitly reduce debug info
size too)
This was worth about 2% for Clang and LLVM - so not a huge win, but a
win. It looks really great for simple STL programs (include <string> and
just declare a string - 14k -> 1.4k of .dwo)
llvm-svn: 202769
Summary:
This merges VFPtrInfo and VBTableInfo into VPtrInfo, since they hold
almost the same information. With that change, the vbtable mangling
code can easily be applied to vftable data and we magically get the
correct, unambiguous vftable names.
Fixes PR17748.
Reviewers: timurrrr, majnemer
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2893
llvm-svn: 202425
Erroring out until we fix the bug means we don't have to keep chasing down
this same miscompile in a bunch of different places.
Differential Revision: http://llvm-reviews.chandlerc.com/D2890
llvm-svn: 202331
Clang is using llvm::StructType::isOpaque() as a way of signaling if
we've finished record type conversion in
CodeGenTypes::isRecordLayoutComplete(). However, Clang was setting the
body of the type before it finished laying out the type as a base type.
Laying out the %class.C.base LLVM type attempts to convert more types,
eventually recursively attempting to layout 'C' again, at which point we
would say that layout was complete, even though we were still in the
middle of it.
By not setting the body, we correctly signal that layout is not
complete, and things work as expected.
At some point, it might be worth refactoring this to avoid looking at
the LLVM IR types under construction.
llvm-svn: 202320
Previously the X86 backend would look for the sret attribute and handle
this for us. inalloca takes that all away, so we have to do the return
ourselves now.
llvm-svn: 202097
Most 64-bit targets define int64_t as long int, and AArch64 should
make same definition to follow LP64 model. In GNU tool chain, int64_t
is defined as long int for 64-bit target. So to get consistent with GNU,
it's better Changing int64_t from 'long long int' to 'long int',
otherwise clang will get different name mangling suffix compared with g++.
llvm-svn: 202004
CGRecordLayoutBuilder was aging, complex, multi-pass, and shows signs of
existing before ASTRecordLayoutBuilder. It redundantly performed many
layout operations that are now performed by ASTRecordLayoutBuilder and
asserted that the results were the same. With the addition of support
for the MS-ABI, such as placement of vbptrs, vtordisps, different
bitfield layout and a variety of other features, CGRecordLayoutBuilder
was growing unwieldy in its redundancy.
This patch re-architects CGRecordLayoutBuilder to not perform any
redundant layout but rather, as directly as possible, lower an
ASTRecordLayout to an llvm::type. The new architecture is significantly
smaller and simpler than the CGRecordLayoutBuilder and contains fewer
ABI-specific code paths. It's also one pass.
The architecture of the new system is described in the comments. For the
most part, the new system simply takes all of the fields and bases from
an ASTRecordLayout, sorts them, inserts padding and dumps a record.
Bitfields, unions and primary virtual bases make this process a bit more
complicated. See the inline comments.
In addition, this patch updates a few lit tests due to the fact that the
new system computes more accurate llvm types than CGRecordLayoutBuilder.
Each change is commented individually in the review.
Differential Revision: http://llvm-reviews.chandlerc.com/D2795
llvm-svn: 201907
Virtual methods expect 'this' to point to the vfptr containing the
virtual method, and this extends to virtual member pointer thunks. The
relevant vfptr is always at offset zero on entry to the thunk, and no
this adjustment is needed.
Previously we would not include the vfptr adjustment in the member
pointer, and we'd look at the vfptr offset when loading from the vftable
in the thunk.
Fixes PR18917.
llvm-svn: 201835
The MS ABI requires that we determine the vbptr offset if have a
virtual inheritance model. Instead, raise an error pointing to the
diagnostic when this happens.
This fixes PR18583.
Differential Revision: http://llvm-reviews.chandlerc.com/D2842
llvm-svn: 201824
In the Microsoft ABI, the vftable is laid out as if all methods in every
overload set were declared in reverse order of declaration at the point
of declaration of the first overload in the set.
Previously we only considered virtual methods in an overload set, but
MSVC includes non-virtual methods for ordering purposes.
Fixes PR18902.
llvm-svn: 201722
Summary:
Generally the vector deleting dtor, which we model as a vtable thunk,
takes care of non-virtual adjustment and delegates to the other
destructor variants. The other non-complete destructor variants assume
that 'this' on entry points to the virtual base subobject that first
declared the virtual destructor.
We need to change the adjustment in both the prologue and the vdtor call
setup.
Reviewers: timurrrr
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2821
llvm-svn: 201612
Extended qualifiers can appear in many places, refactor the code so it's
more reusable. Add tests in areas where we've increased compatibility.
llvm-svn: 201574
Pointer types in the MSVC ABI are a bit awkward, the width of the
pointer is considered a kind of CVR qualifier.
Restrict is handled similarly to const and volatile but is mangled after
the pointer width qualifier.
This fixes PR18880.
llvm-svn: 201569
Previously, we made one traversal of the AST prior to codegen to assign
counters to the ASTs and then propagated the count values during codegen. This
patch now adds a separate AST traversal prior to codegen for the
-fprofile-instr-use option to propagate the count values. The counts are then
saved in a map from which they can be retrieved during codegen.
This new approach has several advantages:
1. It gets rid of a lot of extra PGO-related code that had previously been
added to codegen.
2. It fixes a serious bug. My original implementation (which was mailed to the
list but never committed) used 3 counters for every loop. Justin improved it to
move 2 of those counters into the less-frequently executed breaks and continues,
but that turned out to produce wrong count values in some cases. The solution
requires visiting a loop body before the condition so that the count for the
condition properly includes the break and continue counts. Changing codegen to
visit a loop body first would be a fairly invasive change, but with a separate
AST traversal, it is easy to control the order of traversal. I've added a
testcase (provided by Justin) to make sure this works correctly.
3. It improves the instrumentation overhead, reducing the number of counters for
a loop from 3 to 1. We no longer need dedicated counters for breaks and
continues, since we can just use the propagated count values when visiting
breaks and continues.
To make this work, I needed to make a change to the way we count case
statements, going back to my original approach of not including the fall-through
in the counter values. This was necessary because there isn't always an AST node
that can be used to record the fall-through count. Now case statements are
handled the same as default statements, with the fall-through paths branching
over the counter increments. While I was at it, I also went back to using this
approach for do-loops -- omitting the fall-through count into the loop body
simplifies some of the calculations and make them behave the same as other
loops. Whenever we start using this instrumentation for coverage, we'll need
to add the fall-through counts into the counter values.
llvm-svn: 201528
1. CHECK-NOT-LABEL is not valid FileCheck.
2. This test would not trigger the CHECK-NOT-LABEL even if 'interface' is
replaced with 'struct'.
llvm-svn: 201462
unique them and permits the implementation of dynamic_cast (and
anything else which knows it's working with a complete class
type) to compare their addresses directly.
rdar://16005328
llvm-svn: 201020
Properly support fields that come from anonymous unions and structs
when used as template arguments for pointer to data member params.
llvm-svn: 200921
Properly determine the inheritance model when dealing with nullptr:
- If a nullptr template argument is being checked against
pointer-to-member parameter, nail down an inheritance model.
N.B. We will chose an inheritance model even if we won't ultimately
choose the template to instantiate! Cooky, right?
- Null pointer-to-datamembers have a virtual base table offset of -1,
not zero. Previously, we chose an offset of 0.
llvm-svn: 200920
Function references always use $1? like function pointers and never $E?
like var decl references. Static methods are mangled like function
pointers.
llvm-svn: 200869
Member pointers are mangled as they would be represented at runtime.
They can be a single integer literal, single decl, or a tuple with some
more numbers tossed in. With Clang today, most of those numbers will be
zero because we reject pointers to members of virtual bases.
This change required moving VTableContextBase ownership from
CodeGenVTables to ASTContext, because mangling now depends on vtable
layout.
I also hoisted the inheritance model helpers up to be inline static
methods of MSInheritanceAttr. This makes the AST code that deals with
member pointers much more readable.
MSVC doesn't appear to have stable manglings of null member pointers:
- Null data memptrs in function templates have a mangling collision with
the first field of a non-polymorphic single inheritance class.
- The mangling of null data memptrs changes if you add casts.
- Large null function memptrs in class templates crash MSVC.
Clang uses the class template mangling for null data memptrs and the
function template mangling for null function memptrs to deal with this.
Reviewers: majnemer
Differential Revision: http://llvm-reviews.chandlerc.com/D2695
llvm-svn: 200857
We accept these with a warning in MS mode, but we would previously mark them
invalid, causing us not to emit code for them.
Differential Revision: http://llvm-reviews.chandlerc.com/D2681
llvm-svn: 200815
When a non-trivial parameter is present, clang now gathers up all the
parameters that lack inreg and puts them into a packed struct. MSVC
always aligns each parameter to 4 bytes and no more, so this is a pretty
simple struct to lay out.
On win64, non-trivial records are passed indirectly. Prior to this
change, clang was incorrectly using byval on win64.
I'm able to self-host a working clang with this change and additional
LLVM patches.
Reviewers: rsmith
Differential Revision: http://llvm-reviews.chandlerc.com/D2636
llvm-svn: 200597
This fixes PR15768, where the sret parameter and the 'this' parameter
are in the wrong order.
Instance methods compiled by MSVC never return records in registers,
they always return indirectly through an sret pointer. That sret
pointer always comes after the 'this' parameter, for both __cdecl and
__thiscall methods.
Unfortunately, the same is true for other calling conventions, so we'll
have to change the overall approach here relatively soon.
Reviewers: rsmith
Differential Revision: http://llvm-reviews.chandlerc.com/D2664
llvm-svn: 200587
With this change, we give different results for __alignof than MSVC, but
our record layout is compatible.
Some data member pointers also now have a size that is not a multiple of
their alignment.
Fixes PR18618.
Reviewers: majnemer
Differential Revision: http://llvm-reviews.chandlerc.com/D2669
llvm-svn: 200585
might have a smaller size as compared to the stand-alone type of the base class.
This is possible when the derived class is packed and hence might have smaller
alignment requirement than the base class.
Differential Revision: http://llvm-reviews.chandlerc.com/D2599
llvm-svn: 200031
Some ABIs have different return types for constructors and
destructors, and we're just looking for the end of the function
here. Loosen up the regex.
llvm-svn: 199870
This test requires asserts for now, and exception handling has an
awkward structure that leads to extra run lines. Because of this, the
test file's not a great place for other C++ PGO tests, but
instr-profile.cpp is obviously the better name for them.
llvm-svn: 199863
The MSVC C++ ABI always uses the deduced type in place of auto when
generating external names for variables.
N.B. MSVC doesn't support C++1y's 'operator auto' and this patch will
not give us said functionality.
llvm-svn: 199764
class and use it pervasively to restore debug locations.
Fixes an interaction between cleanup and EH that caused the location
to not be restored properly after emitting a landing pad.
rdar://problem/15208190
llvm-svn: 199444
This makes the C++ ABI depend entirely on the target: MS ABI for -win32 triples,
Itanium otherwise. It's no longer possible to do weird combinations.
To be able to run a test with a specific ABI without constraining it to a
specific triple, new substitutions are added to lit: %itanium_abi_triple and
%ms_abi_triple can be used to get the current target triple adjusted to the
desired ABI. For example, if the test suite is running with the i686-pc-win32
target, %itanium_abi_triple will expand to i686-pc-mingw32.
Differential Revision: http://llvm-reviews.chandlerc.com/D2545
llvm-svn: 199250
Although VBPtrs were being placed correctly by the ms-abi layout engine,
their offsets were being improperly reported to the ASTRecordLayout
builder due to a bug. This patch fixes that and fixes the test cases to
use the correct values.
y
llvm-svn: 199168
Fixes PR18435, where we generated a base ctor instead of a complete
ctor, and so failed to construct virtual bases when constructing the
complete object.
llvm-svn: 199160
In preparation for making the Win32 triple imply MS ABI mode,
make all tests pass in this mode, or make them use the Itanium
mode explicitly.
Differential Revision: http://llvm-reviews.chandlerc.com/D2401
llvm-svn: 199130
The MS abi lays out *all* non-virtual bases with leading vfptrs before
laying out non-virutal bases without vfptrs. This guarantees that the
primary base is laid out first. r198818 fixed RecordLayoutBuilder to
produce compatiable layouts. This patch fixes CGRecordLayoutBuilder to
be able to consume those layouts and produce meaningful output without
tripping any asserts about assumed incoming layout.
A test case is included that shows CGRecordLayoutBuilder in fact
produces output in the compatiable order.
llvm-svn: 198900
issue 1430. Don't allow a pack expansion to be used as an argument to an alias
template unless the corresponding parameter is a parameter pack.
llvm-svn: 198833
This patch refactors microsoft record layout to be more "natural". The
most dominant change is that vbptrs and vfptrs are injected after the
fact. This simplifies the implementation and the math for the offest
for the first base/field after the vbptr.
llvm-svn: 198818
Debug info: Implement a cleaner version of r198461. For symmetry with
C and C++ don't emit an extra lexical scope for the compound statement
that is the body of an Objective-C method.
llvm-svn: 198715
It controls everything that -flimit-debug-info used to, plus the
vtable type optimization. The old -fno-limit-debug-info option is now an
alias to -fstandalone-debug and vice versa.
Standalone is the default on Darwin until dtrace is updated to work with
non-standalone debug info (rdar://problem/15758808).
Note: I kept the LimitedDebugInfo name in CodeGenOptions::DebugInfoKind
because NoStandaloneDebugInfo sounded even more confusing.
llvm-svn: 198655
Summary:
This makes us more compatible with MSVC 2012+ and fixes PR17748 where we
would give two tables the same name.
Rather than doing a fresh depth-first traversal of the inheritance graph
for every record's vbtables, now we memoize vbtable paths for each
record. By doing memoization, we end up considering virtual bases of
subobjects that come later in the depth-first traversal. Where
previously we would have ignored a virtual base that we'd already seen,
we now consider it for name mangling purposes without emitting a
duplicate vbtable for it.
Reviewers: majnemer
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2509
llvm-svn: 198462
Most importantly, this makes our vtable layout match MSVC's. Previously
we would emit a return adjusting thunk whenever the return types
differed, even if the adjustment would have been trivial.
MSVC does emit some trivial return adjusting thunks, but only if there
was already an overridden method that required a return adjustment.
llvm-svn: 198080
These members will still be lazily added to the relevant DWARF DIEs in
LLVM but when enumerating the members they will not appear. This allows
DWARF type units to be more consistent - the type unit will never
contain these special members (so all instances of the type should have
the same DIEs without some having some special members and others having
others) and the special members will be added to the skeletal
declaration that appears in the relevant compile_unit.
llvm-svn: 197844
This was part of the cause for PR17655. We were generating thunks when
we shouldn't have. I suspect that if we tweak the test case for PR17655
to actually require thunks, we can reproduce the same crash.
llvm-svn: 197836
Unlike Itanium's VTTs, the 'most derived' boolean or bitfield is the
last parameter for non-variadic constructors, rather than the second.
For variadic constructors, the 'most derived' parameter comes after the
'this' parameter. This affects constructor calls and constructor decls
in a variety of places.
Reviewers: timurrrr
Differential Revision: http://llvm-reviews.chandlerc.com/D2405
llvm-svn: 197518
clang still doesn't emit the right llvm code when initializing multi-D arrays it seems.
For e.g. the following code would still crash for me on Windows 7, 64 bit:
auto f4 = new int[100][200][300]{{{1,2,3}, {4, 5, 6}}, {{10, 20, 30}}};
It seems that the final new loop that iterates through each outermost array and memsets it to zero gets confused with its final ptr arithmetic.
This patch ensures that it converts the pointer to the allocated type (int [200][300]) before incrementing it (instead of using the base type: 'int').
Richard somewhat squeamishly approved the patch (as a quick fix to potentially make it into 3.4) - while exhorting for a more optimized fix in the future. http://llvm-reviews.chandlerc.com/D2398
Thanks Richard!
llvm-svn: 197294
There's nothing special about type traits accepting two arguments.
This commit eliminates BinaryTypeTraitExpr and switches all related handling
over to TypeTraitExpr.
Also fixes a CodeGen failure with variadic type traits appearing in a
non-constant expression.
The BTT/TT prefix and evaluation code is retained as-is for now but will soon
be further cleaned up.
This is part of the ongoing work to unify type traits.
llvm-svn: 197273
This reverts commit r197184.
Richard Smith brings up some good points, a proper implementation will
require us to mangle unnameable entities compatibly with MSVC.
llvm-svn: 197192
We were mistakengly giving linkonce_odr linkage instead of internal
linkage to the deleting and complete destructor thunks for classes in
anonymous namespaces.
Fixes PR17273.
llvm-svn: 197060
list, each element of the initializer list may provide more than one of the
base elements of the array. Be sure to initialize the right type and bump the
array pointer by the right amount.
llvm-svn: 196995
more than one such initializer in a union, make mem-initializers override
default initializers for other union members, handle anonymous unions with
anonymous struct members better. Fix a couple of semi-related bugs exposed by
the tests for same.
llvm-svn: 196892
With this patch we output the in the order
C2
C1
D2
D1
D0
Which means that a destructor or constructor that call another is output after
the callee. This is a bit easier to read IHMO and a tiny bit more efficient
as we don't put a decl in DeferredDeclsToEmit.
llvm-svn: 196784
Testing has revealed that large integral constants (i.e. > INT64_MAX)
are always mangled as-if they are negative, even in places where it
would not make sense for them to be negative (like non-type template
parameters of type unsigned long long).
To address this, we change the way we model number mangling: always
mangle as-if our number is an int64_t. This should result in correct
results when we have large unsigned numbers.
N.B. Bizarrely, things that are 32-bit displacements like vbptr offsets
are mangled as-if they are unsigned 32-bit numbers. This is a pretty
egregious waste of space, it would be a 4x savings if we could mangle it
like a signed 32-bit number. Instead, we explicitly cast these
displacements to uint32_t and let the mangler proceed.
llvm-svn: 196771
While testing our ability to mangle large constants (PR18175), I
incidentally discovered that we did not properly mangle enums correctly.
Previously, we would append the width of the enum in bytes after the
type-tag differentiator.
This would mean "enum : short" would be mangled as 'W2' while "enum :
char" would be mangled as 'W1'. Upon testing this with several versions
of MSVC, I found that this did not match their behavior: they always use
'W4'.
N.B. Quick testing uncovered that undname allows different numbers to
follow the 'W' in the following way:
'W0' -> "enum char"
'W1' -> "enum unsigned char"
'W2' -> "enum short"
'W3' -> "enum unsigned short"
'W4' -> "enum"
'W5' -> "enum unsigned int"
'W6' -> "enum long"
'W7' -> "enum unsigned long"
However this scheme appears abandoned, I cannot get MSVC to trigger it.
Furthermore, it's incomplete: it doesn't handle "bool" or "long long".
llvm-svn: 196752
Clang outputs LLVM one top level decl at a time. This combined with the
visibility computation code looking for the newest NamespaceDecl would cause
it to produce different results for nested namespaces.
The two options for producing consistent results are
* Delay codegen of anything inside a namespace until the end of the file.
* Don't look for the newest NamespaceDecl.
This patch implements the second option.
This matches the gcc behavior too.
llvm-svn: 196712
This can happen when we're trying to emit a thunk with available_externally
linkage with optimization enabled but bail because it doesn't make sense
for vararg functions.
PR18098.
llvm-svn: 196658
Summary:
MSVC destroys arguments in the callee from left to right. Because C++
objects have to be destroyed in the reverse order of construction, Clang
has to construct arguments from right to left and destroy arguments from
left to right.
This patch fixes the ordering by reversing the order of evaluation of
all call arguments under the MS C++ ABI.
Fixes PR18035.
Reviewers: rsmith
Differential Revision: http://llvm-reviews.chandlerc.com/D2275
llvm-svn: 196402
We would lose track of the mangling number assigned to the original
declaration which would cause us to create manglings that didn't match
the Itanium C++ specification.
e.g. Two static fields with the same name inside of a function template
would receive the same mangling with LLVM fixing up the second field so
they wouldn't collide. This would create an incompatibility with other
compilers following the Itanium ABI.
I've confirmed that the new mangling is identical to the ones generated
by icc and gcc.
N.B. This was uncovered while working on Microsoft mangler.
llvm-svn: 196368
It wasn't possible for an anonymous type to show up inside of function arguments.
However, decltype (which MSVC added support for in 2010) makes this
possible. Further, backrefs to these anonymous types can now be formed.
This fixes PR18022.
N.B. We do not, and very likely _will not_, support MSVC's bug where
subsequent typedefs of anonymous types leak into the linkage name; this
is a gross violation of the ABI. A warning should be introduced to
inform our users of this particular shortcoming.
llvm-svn: 195669
available always-inline functions. This breaks libc++'s locale
implementation. Code generation for this case should be fixed, but this
is a stop gap fix for clang 3.4.
llvm-svn: 195501
This makes Clang emit a linkonce_odr definition for 'val' in the code below,
to be compatible with MSVC-compiled code:
struct Foo {
static const int val = 1;
};
Differential Revision: http://llvm-reviews.chandlerc.com/D2233
llvm-svn: 195283
Instead of storing the vtable offset directly in the function pointer and
doing a branch to check for virtualness at each call site, the MS ABI
generates a thunk for calling the function at a specific vtable offset,
and puts that in the function pointer.
This patch adds support for emitting such thunks. However, it doesn't support
pointers to virtual member functions that are variadic, have an incomplete
aggregate return type or parameter, or are overriding a function in a virtual
base class.
Differential Revision: http://llvm-reviews.chandlerc.com/D2104
llvm-svn: 194827
This patch tests introduces a proper codegen test in place of the
"codegen no longer crashes" test introduced in r193664. The test is also
moved from layout to CodeGenCXX.
Differential Revision: http://llvm-reviews.chandlerc.com/D2174
llvm-svn: 194739
This patch disables aliasing (and rauw) of derived dtors to base dtors at -O0.
This optimization can have a negative impact on the debug quality.
This was a latent bug for some time with local classes, but got noticed when it
was generalized and broke gdb's destrprint.exp.
llvm-svn: 194618
The problem was that given
template<typename T>
struct foo {
~foo() {}
};
template class foo<int>;
We would produce a alias, creating a comdat with D0 and D1, since the symbols
have to be weak. Another TU is not required to have a explicit template
instantiation definition or an explict template instantiation declaration and
for
template<typename T>
struct foo {
~foo() {}
};
foo<int> a;
we would produce a comdat with only one symbol in it.
llvm-svn: 194520
The original decls are created when used. The replacements are created at the
end of the TU in reverse order.
This makes the original order far better for testing. This is particularly
important since the replacement logic could be used even when
-mconstructor-aliases is not used, but that would make many tests hard to read.
This is a fixed version of r194357 which handles replacing a destructor with
another which is an alias to a third one.
llvm-svn: 194452
The assert this patch deletes was valid only when aliasing D2 to D1, not when
looking at a base class. Since the assert was in the path where we had already
decided to not produce an alias, just drop it.
llvm-svn: 194411
The original decls are created when used. The replacements are created at the
end of the TU in reverse order.
This makes the original order far better for testing. This is particularly
important since the replacement logic could be used even when
-mconstructor-aliases is not used, but that would make many tests hard to read.
llvm-svn: 194357
clang-cl adds these, so this makes the tests a bit more realistic. These are the
tests where it would make a difference if the windows specific handling were
removed.
llvm-svn: 194336
It is not safe to emit alias to undefined (not supported by ELF or COFF), but
it is safe to rauw when the alias would have been internal or linkonce_odr.
llvm-svn: 194307
whether we can safely lower a conditional operator to select was insufficient.
I've left a large comment in place to explaining the sort of problems that this
transform can encounter in clang in the hopes of discouraging others from
reimplementing it wrongly again in the future. (The test should also help with
that, but it's easy to work around any single test I might add and think that
your particular implementation doesn't miscompile any code.)
llvm-svn: 194289
Unlike an alias a rauw is always safe, so we don't need to avoid this
optimization when the replacement is not know to be available in every TU.
llvm-svn: 194288
This makes it consistent with -fdump-record-layouts, which was moved to
outs() in r186219. My reasoning for going with stdout is that when one
of these options is present, the layouts are really a program output,
and shouldn't be interleaved with diagnostics, which are on stderr.
Reviewers: timurrrr
Differential Revision: http://llvm-reviews.chandlerc.com/D2127
llvm-svn: 194279
On the microsoft ABI clang is producing one weak_odr and one linkonce_odr
destructor, which is reasonable since only one is required.
The fix is simply to move the assert past the special case treatment of
linkonce_odr.
llvm-svn: 194158
This is a small optimization on linux, but should help more on windows
where msvc only outputs one destructor if there would be two identical ones.
llvm-svn: 194095
deallocation function (and the corresponding unsized deallocation function has
been declared), emit a weak discardable definition of the function that
forwards to the corresponding unsized deallocation.
This allows a C++ standard library implementation to provide both a sized and
an unsized deallocation function, where the unsized one does not just call the
sized one, for instance by putting both in the same object file within an
archive.
llvm-svn: 194055
This is a small optimization on linux, but should help more on windows
where msvc only outputs one destructor if there would be two identical ones.
llvm-svn: 194046
With this patch we produce alias for cases like
template<typename T>
struct foobar {
foobar() {
}
};
template struct foobar<void>;
We just have to be careful to produce the same aliases in every TU because
of comdats.
llvm-svn: 194000
If the sole distinction between two declarations is that one has a
__restrict qualifier then we should not consider it to be an overload.
Instead, we will consider it as an incompatible redeclaration which is
similar to how MSVC, ICC and GCC would handle it.
This fixes PR17786.
N.B. We must not mangle in __restrict into method qualifiers becase we
don't allow overloading between such declarations anymore. To do
otherwise would be a violation of the Itanium ABI.
llvm-svn: 193964
long in linux 64 is 64 bits but is always 32 bits on windows. The lit test was modified
to use long long instead of long and check for 64-bit mangling.
llvm-svn: 193901
Differential Revision: http://llvm-reviews.chandlerc.com/D2082
Adds a lang_c LinkageSpecDecl to lazily generated builtins. This enforces correct
behavior for builtins in a variety of cases without special treatment elsewhere within
the compiler (special treatment is removed by the patch). It also allows for C++
overloads of builtin functions, which Microsoft uses in their headers e.g.
_InterlockedExchangeAdd is an extern C builtin for the long type but an inline wrapper
for int type.
llvm-svn: 193896
The Itanium mangler couldn't cope with mangling an IndirectFieldDecl.
Instead, mangle the field the IndirectFieldDecl refers to.
Further, give IndirectFieldDecl no linkage just like FieldDecl.
N.B. Decl.cpp:getLVForNamespaceScopeDecl tried to calculate linkage for
data members of anonymous structs/unions. However, this seems
impossible so turn it into an assertion.
llvm-svn: 193269
check using the ubsan runtime) and -fsanitize=local-bounds (for the middle-end
check which inserts traps).
Remove -fsanitize=local-bounds from -fsanitize=undefined. It does not produce
useful diagnostics and has false positives (PR17635), and is not a good
compromise position between UBSan's checks and ASan's checks.
Map -fbounds-checking to -fsanitize=local-bounds to restore Clang's historical
behavior for that flag.
llvm-svn: 193205
Summary:
Enforce the rule in C++11 [temp.mem]p2 that local classes cannot have
member templates.
This fixes PR16947.
N.B. C++14 has slightly different wording to afford generic lambdas
declared inside of functions.
Fun fact: Some formulations of local classes with member templates
would cause clang to crash during Itanium mangling, such as the
following:
void outer_mem() {
struct Inner {
template <typename = void>
struct InnerTemplateClass {
static void itc_mem() {}
};
};
Inner::InnerTemplateClass<>::itc_mem();
}
Reviewers: eli.friedman, rsmith, doug.gregor, faisalv
Reviewed By: doug.gregor
CC: cfe-commits, ygao
Differential Revision: http://llvm-reviews.chandlerc.com/D1866
llvm-svn: 193144
This uses function prefix data to store function type information at the
function pointer.
Differential Revision: http://llvm-reviews.chandlerc.com/D1338
llvm-svn: 193058
This removes the dependency on the llvm mangler doing it for us. In isolation,
the benefit is that the testing of what mangling is applied is all in one place:
(C, C++) X (Itanium, Microsoft) are all handled by clang.
This also gives me hope that in the future the llvm mangler (and llvm-ar) will
not depend on TargetMachine.
llvm-svn: 192762
If a class is using the unspecified inheritance model for member
pointers and later we find the class is defined to use single
inheritance, zero out the vbptr offset field of the member pointer when
it is formed.
llvm-svn: 192664
Reid Kleckner