Commit Graph

2407 Commits

Author SHA1 Message Date
Warren Hunt 640015cbc3 [ms-abi] Fixing CGRecordLayoutBuilder w.r.t. MS NonVirutalBase Layout
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
2014-01-09 23:51:31 +00:00
Richard Smith 83b11aae18 PR18401: Fix assert by implementing the current proposed direction of core
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
2014-01-09 02:22:22 +00:00
Warren Hunt d640d7d96e [ms-abi] Refactor Microsoft Record Layout
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
2014-01-09 00:30:56 +00:00
Richard Smith 26d11be738 RP18408: If a member template is used as a template template argument, it does
not cause the template specialization to have no linkage.

llvm-svn: 198726
2014-01-08 01:51:59 +00:00
Adrian Prantl 56741e29d4 Re-applying r198699 after reverting r198461.
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
2014-01-07 22:05:55 +00:00
Adrian Prantl e83b130def Revert "Debug info: Ensure that the last stop point in a function is still within"
This reverts commit r198461.

llvm-svn: 198714
2014-01-07 22:05:52 +00:00
Adrian Prantl a763447124 Implement a new -fstandalone-debug option. rdar://problem/15685848
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
2014-01-07 01:19:08 +00:00
NAKAMURA Takumi f0f82175c8 clang/test/CodeGenCXX/instr-profile.cpp: Appease -Asserts.
FIXME: Don't seek bb labels, like "if.else"
llvm-svn: 198651
2014-01-07 00:59:39 +00:00
Justin Bogner ef512b9929 CodeGen: Initial instrumentation based PGO implementation
llvm-svn: 198640
2014-01-06 22:27:43 +00:00
Reid Kleckner 5f08094106 [ms-cxxabi] Improve vbtable name mangling accuracy
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
2014-01-03 23:42:00 +00:00
Adrian Prantl 96e70d9148 Debug info: Ensure that the last stop point in a function is still within
the lexical block formed by the compound statement that is the function
body.

rdar://problem/15010825

llvm-svn: 198461
2014-01-03 23:34:30 +00:00
Reid Kleckner 604c8b45e4 [ms-cxxabi] Emit fewer trivial return adjusting thunks
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
2013-12-27 19:43:59 +00:00
David Blaikie 3275dc4586 DebugInfo: Do not include implicit members (implicit special members, template instantiations, etc) in the list of members of a structure.
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
2013-12-20 23:19:47 +00:00
Reid Kleckner d2d1cc5156 Compare canonical return types when generating MS C++ ABI vtable thunks
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
2013-12-20 21:44:05 +00:00
Adrian Prantl 0866acd3a3 Debug info: (Bugfix) emit CRV qualifiers for pointers to member functions.
rdar://problem/15678916.

llvm-svn: 197641
2013-12-19 01:38:47 +00:00
Adrian Prantl 0630eb7094 Debug info: Implement (rvalue) reference qualifiers for C++11 non-static
member functions. Paired commit with LLVM.

rdar://problem/15356637

llvm-svn: 197612
2013-12-18 21:48:18 +00:00
Reid Kleckner 89077a1b00 [ms-cxxabi] The 'most derived' ctor parameter usually comes last
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
2013-12-17 19:46:40 +00:00
Hans Wennborg c570061064 Add triple to CodeGenCXX/2004-03-08-ReinterpretCastCopy.cpp in MS ABI mode
I don't think we should need a triple, but this allows us to run the test
until we fix it (PR18251).

llvm-svn: 197326
2013-12-14 17:04:17 +00:00
Alp Toker 82d14cddb6 Remove 'not' from some CodeGen tests
These were just missing an expected-no-diagnostics directive.

Also add -std=c++11 to a test warning noisly about extensions.

llvm-svn: 197308
2013-12-14 04:49:06 +00:00
Faisal Vali 57ae056a5c Quick-Fix pointer arithmetic when performing multi-D new-array initialization.
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
2013-12-14 00:40:05 +00:00
Hans Wennborg 442e4f7f5f Prepare for using MS ABI by default for Win32: update CodeGenCXX tests
llvm-svn: 197281
2013-12-13 22:43:52 +00:00
Alp Toker cbb9034e2a Eliminate BinaryTypeTraitExpr
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
2013-12-13 20:49:58 +00:00
David Majnemer 210e6bfa19 Revert "[-cxx-abi microsoft] Mangle reference temporaries"
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
2013-12-13 00:39:38 +00:00
David Majnemer 0834f13c95 [-cxx-abi microsoft] Mangle reference temporaries
They are mangled the same as normal references, nothing special is going
on here.

llvm-svn: 197184
2013-12-12 23:12:01 +00:00
Rafael Espindola 3497069784 Switch to the new MingW ABI.
GCC 4.7 changed the MingW ABI. On the clang side this means that methods now
have the thiscall calling convention by default.

llvm-svn: 197164
2013-12-12 16:07:11 +00:00
Timur Iskhodzhanov fae1eda671 Add a test case for the PR17748 crash
llvm-svn: 197143
2013-12-12 11:07:03 +00:00
Reid Kleckner a12cd28bb3 [ms-cxxabi] Fix linkage of dtor thunks for anonymous classes
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
2013-12-11 19:21:27 +00:00
Richard Smith 1c96bc5d03 When performing an array new of a multidimensional array with an initializer
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
2013-12-11 01:40:16 +00:00
Richard Smith 0b7f1187aa Add a triple to this IRgen test to unbreak the ARM bots.
llvm-svn: 196953
2013-12-10 19:37:58 +00:00
Richard Smith ab44d5badf Implement DR1460: fix handling of default initializers in unions; don't allow
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
2013-12-10 08:25:00 +00:00
Rafael Espindola c3cde36ead Output destructors and constructors in a more natural order.
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
2013-12-09 14:51:17 +00:00
David Majnemer 2a81645de5 [-cxx-abi microsoft] Mangle large integral constants correctly
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
2013-12-09 10:44:32 +00:00
David Majnemer 048f90cc04 [-cxx-abi microsoft] Properly mangle enums
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
2013-12-09 04:28:34 +00:00
Alp Toker 544bd43624 Remove duplicated -cc1 in tests
llvm-svn: 196728
2013-12-08 18:06:52 +00:00
Rafael Espindola 7ab1ce04bb Fix pr18174.
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
2013-12-08 01:13:22 +00:00
Benjamin Kramer 065c61b646 CodeGen: Don't emit linkage on thunks that aren't emitted because they're vararg.
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
2013-12-07 16:12:52 +00:00
Reid Kleckner ae9451218f [ms-cxxabi] bitcast to i8* to deref a data member pointer
This was causing us to miscompile
llvm::SymbolTableListTraits::getListOwner(), which uses data member
pointers.

llvm-svn: 196545
2013-12-05 22:44:07 +00:00
Alp Toker f6a24ce40f Fix a tranche of comment, test and doc typos
llvm-svn: 196510
2013-12-05 16:25:25 +00:00
Richard Smith f03bd30854 PR17983: Fix crasher bug in C++1y mode when performing a non-global array
delete on a class which has no array cookie and has no class-specific operator
new.

llvm-svn: 196488
2013-12-05 08:30:59 +00:00
Reid Kleckner 739756c0f9 [ms-cxxabi] Construct and destroy call arguments in the correct order
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
2013-12-04 19:23:12 +00:00
David Blaikie 2183e70ce2 Fix test case due to r196394 and improve it to not rely on LLVM code generation either.
llvm-svn: 196399
2013-12-04 19:09:19 +00:00
David Majnemer dbc0c8f941 Sema: Propagate the mangling number into instantiations
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
2013-12-04 09:01:55 +00:00
NAKAMURA Takumi 0acd8a7561 clang/test: REQUIRES: s/x86-64-registered-target/x86-registered-target/
llvm-svn: 196350
2013-12-04 03:41:33 +00:00
Rafael Espindola 623a911672 Check IR on this test.
llvm-svn: 196286
2013-12-03 14:43:17 +00:00
Rafael Espindola 218935d1ec Check IR on this test.
llvm-svn: 196285
2013-12-03 14:37:33 +00:00
Rafael Espindola 6ffd88d664 Check IR on this test.
llvm-svn: 196284
2013-12-03 14:35:42 +00:00
Rafael Espindola 8dcdca67e3 Check IR on this test.
llvm-svn: 196282
2013-12-03 14:33:08 +00:00
Rafael Espindola aec96b866b check IR in this test
llvm-svn: 196281
2013-12-03 14:23:52 +00:00
Rafael Espindola f6b5c16ecc check IR in this test.
llvm-svn: 196280
2013-12-03 14:21:57 +00:00
Rafael Espindola d22e26860d check IR in this test.
llvm-svn: 196279
2013-12-03 14:19:02 +00:00