dc9bc5aeb8
Fix a crash mangling decayed val argument-typed function.
...
// rdar: //8620510 and PR7666
llvm-svn: 118019
2010-11-02 16:54:00 +00:00
86f30f5a1d
Simplify. Builtin types' context is always implied.
...
llvm-svn: 117928
2010-11-01 16:52:40 +00:00
fd88a6160d
Rename getBaseClassOffset to getBaseClassOffsetInBits and introduce a getBaseClassOffset which returns the offset in CharUnits. Do the same thing for getVBaseClassOffset.
...
llvm-svn: 117881
2010-10-31 23:22:37 +00:00
c273f24196
Better solution: calculate the visibility of functions and variables
...
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
2010-10-30 11:50:40 +00:00
7d3dfc0622
Generate bitcasts going in and out of MMX parameters
...
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
2010-10-29 23:12:32 +00:00
e9e32dcb48
Use CodeGenFunction's getContext(), for consistency.
...
llvm-svn: 117734
2010-10-29 22:47:07 +00:00
37bb6c9832
Restore r117644, this time properly ignoring -fvisibility and type visibility
...
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
2010-10-29 22:22:43 +00:00
31d1e2153f
After processing named unions do not fall through to handle anonymous unions.
...
This is tested by funcargs.exp in gdb testsuite.
llvm-svn: 117659
2010-10-29 16:21:19 +00:00
df8792128f
Revert r117644, "Apply visibility in IR gen to variables that are merely
...
declared", it breaks things.
llvm-svn: 117653
2010-10-29 15:19:36 +00:00
077dc60540
Don't assert on attempts to throw 'bool'. I wonder if in the history of C++
...
anyone has ever intentionally done this outside of a compiler test case.
llvm-svn: 117645
2010-10-29 08:14:02 +00:00
4af6bf1fdc
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: 117644
2010-10-29 07:49:41 +00:00
ed23f18d7e
Basic types are language defined builtins. They are always defined at top most level.
...
llvm-svn: 117613
2010-10-28 22:03:20 +00:00
565141612f
When emitting l-values for bool non-__block decl references, make a pointer
...
using the memory type; fixes an assert.
Fixes rdar://problem/8605032
llvm-svn: 117610
2010-10-28 21:37:57 +00:00
faadd7b72d
Directly use NamespaceDecl->getLocation() to find the source file.
...
llvm-svn: 117576
2010-10-28 19:12:46 +00:00
8e0073008a
Stay within 80 columns.
...
llvm-svn: 117561
2010-10-28 17:27:32 +00:00
9de9160d55
Implement an indirect-goto optimization for goto *&&lbl and respect this
...
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
2010-10-28 08:53:48 +00:00
1bee63f325
Fix context info for enums.
...
Radar 8595129
llvm-svn: 117507
2010-10-27 23:23:58 +00:00
3a7f6926d1
Restore r117403 (fixing IR gen for bool atomics), this time being less
...
aggressive about the form we expect bools to be in. I don't really have
time to fix all the sources right now.
llvm-svn: 117486
2010-10-27 20:58:56 +00:00
9d798a07e4
Revert r117403 as it caused PR8480.
...
llvm-svn: 117456
2010-10-27 17:13:49 +00:00
1518a5eca7
Do the guarding of instantiated static data members
...
on if its linkage is weak. Currently this is the
case but may change in the future. (part of radar
8562966).
llvm-svn: 117452
2010-10-27 16:21:54 +00:00
b00c2144b3
Also devirtualize calls to a member functions where the containing class has been marked final.
...
llvm-svn: 117445
2010-10-27 13:34:43 +00:00
a7911fa3d7
If a virtual member function has the 'final' attribute, we can devirtualize calls to it.
...
llvm-svn: 117444
2010-10-27 13:28:46 +00:00
67ca8c4c7b
Patch to provide guard when initializing instances
...
of static data member of a class template.
Fixes //rdar :// 8562966 and pr8409.
llvm-svn: 117410
2010-10-26 22:47:47 +00:00
6bde954f47
Extract procedures to do scalar-to-memory and memory-to-scalar conversions
...
in IR gen, and use those to fix a correctness issue with bool atomic
intrinsics. rdar://problem/8461234
llvm-svn: 117403
2010-10-26 22:09:15 +00:00
8fc50c296a
Factor out the code for emitting code to load vtable pointer members
...
so that it's done in one place.
llvm-svn: 117386
2010-10-26 18:44:08 +00:00
dc2c4db598
Add infrastructure for emitting TBAA metadata with the "constant" flag.
...
llvm-svn: 117328
2010-10-25 23:51:23 +00:00
d831356635
Trim an unnecessary #include.
...
llvm-svn: 117321
2010-10-25 21:59:29 +00:00
7dfd13c167
Factor out the code for creating the Root and Char nodes, so that
...
they can be used outside of the main getTBAAInfo function.
llvm-svn: 117320
2010-10-25 21:48:30 +00:00
c123623d28
After discussion with Doug and John, I am reverting
...
the patch.
llvm-svn: 117159
2010-10-22 22:05:03 +00:00
457a04e3ce
Substantially revise how clang computes the visibility of a declaration to
...
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
2010-10-22 21:05:15 +00:00
0196aa28ec
Patch fixes miscompile with non-trivial copy constructors and
...
statement expressions, //rdar: //8540501
llvm-svn: 117146
2010-10-22 21:01:02 +00:00
401c916c8a
Delay record type's debug info emission, in -flimit-debug-info mode, if member expression's base is call expr.
...
llvm-svn: 117127
2010-10-22 18:56:27 +00:00
db3910000a
Revert unintentional check-in.
...
llvm-svn: 117120
2010-10-22 18:31:12 +00:00
f79199d140
Tidy up MIPS_linkage name. Provide it only if it does not match regular name, otherwise it confuses debugger.
...
This is tested by local.C in llvmgcc testsuite.
llvm-svn: 117107
2010-10-22 17:11:50 +00:00
fb5e584151
More class anonymization.
...
llvm-svn: 117106
2010-10-22 16:48:22 +00:00
2bb62c3d2b
Rename clang's TBAA tree root from "experimental" to "simple".
...
llvm-svn: 117046
2010-10-21 18:50:04 +00:00
5419ce6827
Add some more comments.
...
llvm-svn: 117043
2010-10-21 18:49:12 +00:00
50c925fe96
This patch implements Next's IRGen for -fconstant-string-class=class-name.
...
PR6056, //rdar: //8564463
llvm-svn: 116819
2010-10-19 17:19:29 +00:00
f5a1fbcdf3
Fix Whitespace.
...
llvm-svn: 116798
2010-10-19 06:39:39 +00:00
9987c0ea42
We shouldn't keep track of MMX registers "needed" separately from the SSE
...
registers needed.
llvm-svn: 116772
2010-10-18 23:51:38 +00:00
18e4edae82
Fix some bugs in local class mangling brought up in PR8355.
...
Patch by Richard Smith!
llvm-svn: 116752
2010-10-18 21:28:44 +00:00
f34011e4ca
patch fixes class names missing from method names in debug information for
...
synthesized property. // rdar: //8498026
llvm-svn: 116717
2010-10-18 17:51:06 +00:00
5cd41c4b13
Reapply r116684 with fixes. The test cases needed to be updated.
...
llvm-svn: 116696
2010-10-18 03:41:31 +00:00
c7c9be661f
Temporarily revert r116684. It was causing failures with
...
Clang :: CodeGen/x86_32-arguments-darwin.c
Clang :: CodeGen/x86_32-arguments-linux.c
llvm-svn: 116687
2010-10-17 07:58:46 +00:00
812f4b123e
The "gcc.dg/compat/vector-1 -m32" test was broken after the MMX rewrite. The
...
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
2010-10-17 07:38:01 +00:00
d509182336
Coding by inspection has its problems.
...
llvm-svn: 116672
2010-10-16 16:34:08 +00:00
17afe45a87
objc_exception_rethrow does not take an exception argument.
...
rdar://problem/8535238
llvm-svn: 116663
2010-10-16 08:21:07 +00:00
622581b73b
Revert r116656, "IRgen/Obj-C/NeXT: Fix the IR signature for
...
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
2010-10-16 05:04:10 +00:00
90bb5c2315
IRgen/Obj-C/NeXT: Fix the IR signature for objc_exception_rethrow, so we don't
...
generate unnecessary %al clear on x86_64.
llvm-svn: 116656
2010-10-16 04:08:16 +00:00
f47df3e02a
Mention the paragraphs in the standards which sanction TBAA.
...
llvm-svn: 116617
2010-10-15 20:54:41 +00:00
c44fd6486e
Use a different name for pointer types in tbaa, to be a little
...
more consistent with other names, and to look less like a magic name.
llvm-svn: 116616
2010-10-15 20:26:20 +00:00
2d0a3c7b8c
Add a comment.
...
llvm-svn: 116615
2010-10-15 20:24:53 +00:00
4a3b1b3ef4
Fix a typo in a comment Frits van Bommel spotted.
...
llvm-svn: 116614
2010-10-15 20:24:10 +00:00
2e29eb5a4b
Experimental TBAA support for enum types.
...
llvm-svn: 116613
2010-10-15 20:23:12 +00:00
d19ee8a07f
Don't leak the TBAA object.
...
llvm-svn: 116595
2010-10-15 18:04:46 +00:00
f5c5e0773e
Add a comment about odd "signed char" incompatibility between C++ and C.
...
llvm-svn: 116593
2010-10-15 17:52:03 +00:00
1c9c3fd50a
Death to blocks, or at least the word "block" in one particular obnoxiously
...
ambiguous context.
llvm-svn: 116567
2010-10-15 04:57:14 +00:00
59ae137329
Update CMake.
...
llvm-svn: 116548
2010-10-15 00:39:31 +00:00
d65c196439
Add a TBAA type for pointer types.
...
llvm-svn: 116544
2010-10-15 00:01:39 +00:00
3f1cf0f373
Unsigned types are TBAA-compatible with their signed counterparts.
...
Also, handle unknown types conservatively.
llvm-svn: 116541
2010-10-14 23:39:00 +00:00
947c9af774
Experimental TBAA support.
...
This enables metadata generation by default, however the TBAA pass
in the optimizer is still disabled for now.
llvm-svn: 116536
2010-10-14 23:06:10 +00:00
0f58561907
Use root non-virtual primary base class, not just immediate primary base class, for AT_containing_type.
...
This is tested by virtfunc.exp in gdb testsuite.
llvm-svn: 116535
2010-10-14 22:59:23 +00:00
681c0754d9
Eliminate usage of ObjCSuperExpr used for
...
'super' as receiver of property or a setter/getter
methods. //rdar: //8525788
llvm-svn: 116483
2010-10-14 16:04:05 +00:00
67f70aaf5a
Fix debug info for anon union.
...
This is tested by anon-union.exp.
llvm-svn: 116372
2010-10-12 23:24:54 +00:00
77890879dd
Fixes a typo which caused byte offset in debug info
...
for ivars to be 0. Fixes pr8353.
llvm-svn: 116273
2010-10-11 23:55:47 +00:00
33ddf69fb3
Fix debug info for functions whose context is a namespace.
...
This is tested by namespace.exp in gdb testsuite.
llvm-svn: 116248
2010-10-11 21:58:41 +00:00
a09e8efd1f
Per discussion with Sanjiv, remove the PIC16 target from mainline. When/if
...
it comes back, it will be largely a rewrite, so keeping the old codebase
in tree isn't helping anyone.
llvm-svn: 116191
2010-10-11 05:44:49 +00:00
0c34b138df
Make sure the VTables for template instantiations are emitted even if the key function doesn't have a body.
...
llvm-svn: 116186
2010-10-11 03:25:57 +00:00
f0e31c8b9c
Don't let typeinfo name symbols be 'internal hidden', it can lead to linker conflicts
...
with similarly named classes in anonymous namespaces.
llvm-svn: 116185
2010-10-11 03:25:53 +00:00
11f6be1ce8
lib/CodeGen/CodeGenModule.cpp: DLLImportLinkage should be processed also on declaration.
...
It enables clang to compile Mingw's headers.
llvm-svn: 116184
2010-10-11 02:30:56 +00:00
073c9cb592
Implement __builtin_ia32_vec_ext_v2si function (required by Qt).
...
llvm-svn: 116162
2010-10-10 03:19:11 +00:00
65b2a965fb
Add target implementations for the X86 builtins:
...
__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
2010-10-09 08:47:25 +00:00
a2fabff4f6
Permit constant evaluation of const floating-point variables with
...
constant initializers.
llvm-svn: 116138
2010-10-09 01:34:31 +00:00
0bf3140424
Implement C++0x scoped enumerations, from Daniel Wallin! (and tweaked a
...
bit by me).
llvm-svn: 116122
2010-10-08 23:50:27 +00:00
7a2b425ee1
Revert the hack Chris Lattner added in r97981 to work around
...
brokenness in the designated-initializer ASTs. The ASTs were fixed by
Alp Toker's patch (r116098) for PR6955. Fixes PR6537.
llvm-svn: 116101
2010-10-08 20:56:19 +00:00
3db5480b27
Tighter check in r116060 blcoked enums also. Emit enum const's debug info.
...
llvm-svn: 116071
2010-10-08 18:24:19 +00:00
5e8ee2bb89
Do not repeat debug info for file variable constants.
...
This is tested by file-statics.exp in gdb testsuite.
llvm-svn: 116060
2010-10-08 17:02:40 +00:00
251f859b23
Identify functions with prototype appropriately in debug info.
...
This is tested by callfuncs.exp in gdb testsuite.
llvm-svn: 115989
2010-10-07 22:03:49 +00:00
284fa417ec
Mark FunctionNoProtoType's argument as unspecified parameters.
...
llvm-svn: 115834
2010-10-06 20:51:45 +00:00
569eafce63
Re-enable EH cleanups to destroy __block variables, now that we have a moment to
...
deal with the consequences. Fixes rdar://problem/8224178.
llvm-svn: 115816
2010-10-06 18:56:43 +00:00
c0601d1e99
Do not emit subrange for incomplete array type.
...
This is tested by ptype.exp in gdb testsuite.
llvm-svn: 115805
2010-10-06 18:30:00 +00:00
5b389f495b
Remove unused argument.
...
llvm-svn: 115789
2010-10-06 15:58:57 +00:00
e5dd32da11
Teach PopCleanupBlock to correctly handle the possibility of branching through
...
a EH-only cleanup as part of a fallthrough branch-through. That this happens
for this test case is actually a separate bug.
llvm-svn: 115668
2010-10-05 20:48:15 +00:00
0077b22aa4
If we're resolving all outstanding fixups, and there are multiple fixups
...
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
2010-10-05 02:33:56 +00:00
9916e3fa93
In the fragile ObjC ABI, save the caught exception to the side if there are
...
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
2010-10-04 23:42:51 +00:00
3703ff4163
dyn_cast is more appropriate here.
...
llvm-svn: 115569
2010-10-04 22:28:23 +00:00
95eea45aa5
Add missing '}' :)
...
llvm-svn: 115568
2010-10-04 22:13:18 +00:00
44b8bf0f5b
Emit debug info for an aggregate while processing MemberExpr if the aggregate's debug info was delayed untill now.
...
llvm-svn: 115564
2010-10-04 21:46:04 +00:00
cbc7bc94c9
Fix filename in header comment.
...
llvm-svn: 115561
2010-10-04 21:15:33 +00:00
64d7f2a014
when expanding a builtin, if the argument is required to be a constant,
...
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
2010-10-02 00:09:12 +00:00
07e96866a2
tidy
...
llvm-svn: 115383
2010-10-01 23:43:16 +00:00
d18c5aa20a
Mark explict methods as explict in debug info.
...
llvm-svn: 115379
2010-10-01 23:32:17 +00:00
885e9dff93
Output debug info. for ivars declared in class
...
extension and implementation.
Fixes rdar://8493239.
llvm-svn: 115248
2010-10-01 00:01:53 +00:00
91bbb5547d
Introduce -flimit-debug-info.
...
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
2010-09-30 19:05:55 +00:00
7c086227e6
Attach aritifical attribute with implicit parameters.
...
Radar 8493141.
llvm-svn: 115104
2010-09-29 23:09:21 +00:00
330b65e350
Emit method access specifier.
...
Radar 8490416.
llvm-svn: 115090
2010-09-29 21:46:16 +00:00
db2732ac42
Update to reflect DIFactory::CreateSubprogram() interface change from r115084.
...
llvm-svn: 115085
2010-09-29 21:05:52 +00:00
984cdeef37
Remove dead code.
...
llvm-svn: 115079
2010-09-29 20:05:01 +00:00
8caf6415a8
Add support for attribute((naked)), patch by Zoxc on cfe-commits!
...
- Minor style tweaks by me.
llvm-svn: 115056
2010-09-29 18:20:25 +00:00
11191f11b8
Accidentally committed some temporary changes on my branch when reverting patches.
...
llvm-svn: 114936
2010-09-28 01:28:56 +00:00
6d8c442e08
Temporarily revert 114929 114925 114924 114921. It looked like they (or at least
...
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
2010-09-28 01:09:49 +00:00
ec3bec0c7a
Kill FunctionDecl's IsCopyAssignment bit; it duplicated what could
...
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
2010-09-27 22:37:28 +00:00
1308667f18
Revert my patch changing the MMX "shift" intrinsics that take immediates into
...
"shift with non-immediate" intrinsics. It gets here because we they aren't
immediates anymore.
llvm-svn: 114890
2010-09-27 21:22:25 +00:00
8fb87aec78
Patch implements passing arrays to functions expecting
...
vla. Implements pr7827.
llvm-svn: 114737
2010-09-24 17:30:16 +00:00
19964dbe3b
IRgen/ABI/ARM: Return large vectors in memory.
...
llvm-svn: 114619
2010-09-23 01:54:32 +00:00
b34b08098c
IRgen/ABI/ARM: Trust the backend to pass vectors correctly for the given ABI.
...
- 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
2010-09-23 01:54:28 +00:00
b2f659b7a0
fix the rest of rdar://8461279 - clang miscompiles address-space qualified atomics
...
llvm-svn: 114503
2010-09-21 23:40:48 +00:00
c9066d3072
same bug as before, this time with __sync_val_compare_and_swap.
...
llvm-svn: 114502
2010-09-21 23:35:30 +00:00
7cf46bfda0
fix __sync_bool_compare_and_swap to work with address-space qualified types.
...
llvm-svn: 114498
2010-09-21 23:24:52 +00:00
521c72c756
Fixes an IRgen ICE due to cast of null pointer to
...
a vla type (fixes pr7827).
llvm-svn: 114495
2010-09-21 22:53:33 +00:00
cb75021034
IRgen for gnu extension's conditional lvalue expression
...
with missing LHS. radar 8453812. Executable test is checked
into llvm test suite.
llvm-svn: 114457
2010-09-21 18:32:21 +00:00
3e0c527dcc
Adjust for debug info API change.
...
llvm-svn: 114438
2010-09-21 15:59:59 +00:00
8162d4ad31
Implements in IRgen gnu extensions missing LHS for
...
complex conditionals. Radar 8453812.
llvm-svn: 114376
2010-09-20 23:50:22 +00:00
abb5a7370d
Check in support for OpenCL conditional operator on vector types.
...
llvm-svn: 114371
2010-09-20 22:41:17 +00:00
4fbf4af0f6
Fixes IRgen bug in objc++ reference binding of a
...
getter expression.
Fixes // rdar://8437240
llvm-svn: 114299
2010-09-18 20:47:25 +00:00
bf98608a3d
do not rely on the implicit-dereference semantics of dyn_cast_or_null
...
llvm-svn: 114281
2010-09-18 13:00:17 +00:00
f09d96f76d
Adjust a fixup's starting branch if it's being resolved because
...
it reached the outermost scope and it hasn't yet been forwarded
to a cleanup. Fixed PR8175.
llvm-svn: 114259
2010-09-18 02:24:39 +00:00
1224061281
Added '|' delimiter to separate inline asm multiple alternative constraints for Clang side of support.
...
llvm-svn: 114253
2010-09-18 01:15:13 +00:00
8ea46b6654
Fix a bug with binding l-values to elided temporaries, and leave a couple
...
helpful asserts behind.
llvm-svn: 114250
2010-09-18 00:58:34 +00:00
d632616f86
The MMX shift-with-immediate builtins require the equivalent
...
shift-with-immediate LLVM intrinsics.
llvm-svn: 114239
2010-09-17 23:46:16 +00:00
dd84ef1e62
Add a -ftrapv-handler= option which allows a handler to invoke instead of simply aborting when a signed operation overflows. This mirrors the (GCC-incompatible) behaviour from clang 1.0 and 1.1 when -ftrapv was specified, but allows the handler to be defined for each compilation unit.
...
llvm-svn: 114192
2010-09-17 18:29:54 +00:00
5bbd1b0051
Patch to add IRgen support for Gnu's conditional operator
...
extension when missing LHS. This patch covers scalar
conditionals only. Others are wip.
(pr7726, radar 8353567).
llvm-svn: 114182
2010-09-17 15:51:28 +00:00
195fa00399
IRgen: Change CodeGenPasses to be a PassManager, so it can have CallGraphSCC or
...
Module. Patch by Mike Gist!
llvm-svn: 114171
2010-09-17 07:35:16 +00:00
769250ea70
Currently we're initializing the vtable pointers of a class only after
...
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
2010-09-17 02:31:44 +00:00
7f9c92a9a0
When emitting a new-expression inside a conditional expression,
...
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
2010-09-17 00:50:28 +00:00
dd38fbc7fb
IRgen/ABI/x86-32: Realign indirect arguments when the ABI requires us to pass
...
them with a smaller alignment than the rest of codegen expects.
llvm-svn: 114115
2010-09-16 20:42:06 +00:00
7b7c2937ef
IRgen/ABI: Add support for realigning structures which are passed by indirect
...
reference.
llvm-svn: 114114
2010-09-16 20:42:02 +00:00
ed23de3348
IRgen/ABI/x86_32/Darwin: On Darwin, only structures with SSE vector types get passed
...
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
2010-09-16 20:42:00 +00:00
8a6c91ff76
IRgen/x86_32/Linux: Linux seems to align all stack objects to 4 bytes, unlike
...
Darwin. Checked vs the handiest Linux llvm-gcc I had around, someone on Linux is
welcome to investigate more.
llvm-svn: 114112
2010-09-16 20:41:56 +00:00
c58ec3a9e5
Removes unused setter.
...
llvm-svn: 114085
2010-09-16 16:14:34 +00:00
0bdb1fd477
Opportunistically use the C++ personality function in ObjC++
...
translation units that don't catch ObjC types. rdar://problem/8434851
llvm-svn: 114070
2010-09-16 06:16:50 +00:00
2a8b9a3caa
Further beautification: this pointer can now be typed.
...
llvm-svn: 114061
2010-09-16 03:16:41 +00:00
58649dc609
Initialize AggValueSlot's flags along all paths, plus minor beautification.
...
Prospective fix for broken commit in r114045.
llvm-svn: 114060
2010-09-16 03:13:23 +00:00
b60e70f963
Patch to move RequiresGCollection bit to
...
AggValueSlot slot.
llvm-svn: 114045
2010-09-16 00:20:07 +00:00
28b5286bda
While handling change of file, check if _current_ file is already seen or not. If current file is seen then it indicates that end of previous file's lexical scope.
...
This fixes radar 8396182.
llvm-svn: 114018
2010-09-15 20:50:40 +00:00
7a626f63f7
one piece of code is responsible for the lifetime of every aggregate
...
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
2010-09-15 10:14:12 +00:00
d441b1e637
Tweak this assert.
...
llvm-svn: 113863
2010-09-14 21:45:42 +00:00
824c2f537c
Implement the EH cleanup to call 'operator delete' if a new-expression throws
...
(but not if destructors associated with the full-expression throw).
llvm-svn: 113836
2010-09-14 07:57:04 +00:00
9efa1ce145
Fix VLA miscompilation.
...
llvm.stacksave/llvm.stackrestore wasn't emitted for VLAs in inner scopes.
Fixes r8403108.
llvm-svn: 113822
2010-09-14 00:42:34 +00:00
09476212de
Revert "CMake: Update to use standard CMake dependency tracking facilities instead"
...
This reverts commit r113631
Conflicts:
CMakeLists.txt
lib/CodeGen/CMakeLists.txt
llvm-svn: 113817
2010-09-13 23:54:41 +00:00
8eb06f17c4
Eagerly evaluate type traits in Sema instead of lazily in AST. They actually need Sema access to be correct, fixes coming up.
...
llvm-svn: 113782
2010-09-13 20:56:31 +00:00
0aa35b9cf5
Block description for trivial block literals have
...
their 'isa' field scanned regardless.
llvm-svn: 113749
2010-09-13 16:09:44 +00:00
8ee6a4142a
Augment -ftrapv to check for divide by zero, mod by zero, and INT_MIN % -1.
...
Patch by John Regehr!
llvm-svn: 113705
2010-09-11 21:47:09 +00:00
9efc54890d
CMake: Fix mingw32 build.
...
llvm-svn: 113677
2010-09-11 02:13:48 +00:00
933c6723a4
Fixes an obscure bug in importd block variable layout
...
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
2010-09-11 01:27:29 +00:00
0881f4a367
CMake: Update to use standard CMake dependency tracking facilities instead
...
of whatever we were using before...
llvm-svn: 113631
2010-09-10 21:13:16 +00:00
b67655fc6d
Implement CXXNoexceptExpr codegen.
...
llvm-svn: 113629
2010-09-10 21:04:00 +00:00
4202c0f2a9
Define and implement CXXNoexceptExpr. Create it in Sema.
...
llvm-svn: 113623
2010-09-10 20:55:43 +00:00
f93ac89414
IRGen fix for using property-dot syntax to pass
...
reference object to a c++ member function.
fixes radar 8409336.
llvm-svn: 113602
2010-09-10 18:56:35 +00:00
c9199f522f
Clean up CMake dependencies
...
llvm-svn: 113489
2010-09-09 15:44:58 +00:00
cfddabf5a3
Block ivar layout must assume that the 'isa'
...
field of the block descriptor is GC'able (scanned)
as this what the runtime expects (one can send it
messages). Radar 8394947.
llvm-svn: 113454
2010-09-09 00:21:45 +00:00
5cc0a67f08
Fix warnings caused by new CXXUuidofExprClass enumerator.
...
llvm-svn: 113444
2010-09-08 23:47:05 +00:00
68ff03728a
Implement ARM static local initialization guards, which are more compact than
...
Itanium guards and use a slightly different compiled-in API.
llvm-svn: 113330
2010-09-08 01:44:27 +00:00
366a94822b
Local static block variable referecned in its
...
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
2010-09-07 23:26:17 +00:00
535618b927
get rid of a warning.
...
llvm-svn: 113256
2010-09-07 19:57:04 +00:00
56603ef7b2
Have Sema check for validity of CGString literal
...
instead of asserting in IRGen. Fixes radar 8390459.
llvm-svn: 113253
2010-09-07 19:38:13 +00:00
2fdb5b5955
LastFieldBitfield in CGObjCCommonMac::BuildAggrIvarLayout keeps bitfields or unnamed fields but later the code
...
assumes that it's always a bitfield. This can lead to a crash (reported at rdar://8368320).
llvm-svn: 113154
2010-09-06 12:00:10 +00:00
52bcf96384
move the hackaround for PR6537 to catch unions as well,
...
fixing the ICE in PR7151
llvm-svn: 113130
2010-09-06 00:13:11 +00:00
f53c096813
clean up some formatting.
...
llvm-svn: 113129
2010-09-06 00:11:41 +00:00
1b36b55f89
No functional change. Replace Out << 'a' << 'b' with Out << "ab" and spell
...
David Vandevoorde's name correctly.
llvm-svn: 113103
2010-09-05 03:40:33 +00:00
60babfb8a1
Casting of a property reference to 'void' did not
...
generate the necessary code. This patch fixes it.
// rdar://8389655
llvm-svn: 113079
2010-09-04 19:49:18 +00:00
2917bd0fb4
Petty optimization.
...
llvm-svn: 113049
2010-09-04 01:26:37 +00:00
087206dbcd
Truncate block variable of bool type to i1 when its
...
value is used. This matches with non-block variable
use of bool type. (Fixes radar 8390062).
llvm-svn: 113027
2010-09-03 23:07:53 +00:00
9d798d13f3
Cope with llvm's reference to bool type of 'i1' vs. clang's
...
type of 'i8' for the same for __block variables of
type bool. refixes radar 8382559.
llvm-svn: 113015
2010-09-03 21:36:02 +00:00
56f57589af
A constant initializer never matches the type of the variable it's
...
initializing; it at best matches the element type of the variable
it's initializing. Fixes PR8073.
llvm-svn: 112992
2010-09-03 18:58:50 +00:00
2f8df98c92
IRgen: Fix silly thinko in r112021, which was generating code for the same expr
...
twice. This showed up as an assert on the odd test case because we generated the
decl map entry twice.
llvm-svn: 112943
2010-09-03 02:07:00 +00:00
335e16bad4
Add symantic support for the Pascal calling convention via
...
"__attribute((pascal))" or "__pascal" (and "_pascal" under
-fborland-extensions). Support still needs to be added to llvm.
llvm-svn: 112939
2010-09-03 01:29:35 +00:00
6061b05d51
Translate NEON vabdl, vaba, and vabal builtins to be implemented using the
...
vabd intrinsic combined with zext and add operations.
llvm-svn: 112937
2010-09-03 01:27:09 +00:00
0d635f53a8
Re-commit r112916 with an additional fix for the self-host failures.
...
I've audited the remaining getFunctionInfo call sites.
llvm-svn: 112936
2010-09-03 01:26:39 +00:00
c32f94b4ce
Revert r112916, it's breaking selfhost pretty badly.
...
llvm-svn: 112925
2010-09-03 00:40:45 +00:00
be13194655
Revert "Another i1 vs. i8 type mismatch issue. This time", it breaks some projects.
...
llvm-svn: 112922
2010-09-03 00:35:23 +00:00
12d3891a27
It's not safe to use the generic CXXMethodDecl overload of CGT::getFunctionInfo
...
to set up a destructor call, because ABIs can tweak these conventions.
Fixes rdar://problem/8386802.
llvm-svn: 112916
2010-09-03 00:01:57 +00:00
a70fab8dd7
IRgen/Obj-C: Rewrite Objective-C bit-field access to compute the access strategy
...
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
2010-09-02 23:53:31 +00:00
c7f9bbafe4
IRgen: Move CGBitFieldInfo strategy computation helpers to static member
...
functions.
llvm-svn: 112913
2010-09-02 23:53:28 +00:00
040ad500c4
Fix a few more ConvertTypes that should be ConvertTypeForMems, fixing
...
two regressions in Boost.Config.
llvm-svn: 112908
2010-09-02 23:24:14 +00:00
5b4904f7a3
Add a bunch of missing bitcasts for clang NEON builtin expansions.
...
Radar 8388233
2010-09-02 22:37:30 +00:00
6429f5cf29
Eliminate CXXBindReferenceExpr, which was used in a ton of
...
well-intentioned but completely unused code.
llvm-svn: 112868
2010-09-02 21:50:02 +00:00
cf60744cd7
Tidy up last commit, as per Devang's comments.
...
llvm-svn: 112840
2010-09-02 18:01:51 +00:00
04f3621860
Revert my two IRgen fixes for "bool", then use a far simpler approach
...
based on ConvertTypeForMem. Thanks to John for pointing out the right
solution.
llvm-svn: 112838
2010-09-02 17:38:50 +00:00
05bbef63b6
Another i1 vs. i8 type mismatch issue. This time
...
a 'bool' byref variable in memory. Fixes radar 8382559.
llvm-svn: 112835
2010-09-02 17:28:31 +00:00
6bf98ff491
Use the unmangled name for the display name in Objective-C debug info. This should have no effect with the Mac runtime where clang (unlike GCC) uses the display name symbol name.
...
llvm-svn: 112833
2010-09-02 17:16:32 +00:00
262e4e2ab5
Fix more i1/i8 pointer madness. Here, an overactive assertion
...
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
2010-09-02 15:34:35 +00:00
e791a0546c
Fix a crash involving pointer-to-data-members of boolean type. We were
...
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
2010-09-02 15:00:29 +00:00
9bca923c10
Simplify some random accesses to the ASTContext from the ABI classes.
...
llvm-svn: 112816
2010-09-02 10:25:57 +00:00
8ed55a54fd
Abstract IR generation of array cookies into the C++ ABI class and
...
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
2010-09-02 09:58:18 +00:00
1b87c9a646
Translate NEON vmull, vmlal, and vmlsl builtins to llvm multiply-add/sub
...
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
2010-09-01 23:20:27 +00:00
61a312413c
Fix IRGen when property-dot syntax used to access
...
a c++ class object 'ivar'. Fixes radar 8366604.
llvm-svn: 112729
2010-09-01 19:36:41 +00:00
0ef508d301
Split ObjCInterfaceDecl::ReferencedProtocols into two lists: ReferencedProtocols and AllReferencedProtocols. ReferencedProtocols
...
(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
2010-09-01 01:21:15 +00:00
d426c8eae3
fix rdar://8360877 a really nasty miscompilation in Boost.Xpressive
...
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
2010-09-01 00:50:20 +00:00
be5eb17536
same refactoring as before, this time on the argument side.
...
llvm-svn: 112684
2010-09-01 00:24:35 +00:00
52b3c13149
refactor some code to cut down on redundancy, no functionality change.
...
llvm-svn: 112683
2010-09-01 00:20:33 +00:00
c134eb5ada
Amusingly, I missed this point of abstraction in all my earlier
...
member-pointer refactoring: dereferencing a member data pointer.
llvm-svn: 112640
2010-08-31 21:07:20 +00:00
04dc957260
Add support for windows x86-64 varargs, patch by Cameron Esfahani!
...
llvm-svn: 112603
2010-08-31 16:44:54 +00:00
5d865c3292
Teach IR generation to return 'this' from constructors and destructors
...
under the ARM ABI.
llvm-svn: 112588
2010-08-31 07:33:07 +00:00
50c682585f
Rename DeclContext::getLookupContext to getRedeclContext and change its semantics slightly. No functionality change in the absence of inline namespaces. Also, change a few places where inline namespaces actually make a difference to be prepared for them.
...
llvm-svn: 112563
2010-08-31 00:36:30 +00:00
a48fbe8c53
Fix PR8029, a x86-32 ABI regression in introduced in r112211
...
llvm-svn: 112537
2010-08-30 22:03:23 +00:00
b9225f7f85
Translate NEON vmovn builtin to a vector truncation instead of using an llvm
...
intrinsic.
llvm-svn: 112504
2010-08-30 19:57:13 +00:00
3b7918625c
Revert my user-defined literal commits - r1124{58,60,67} pending
...
some issues being sorted out.
llvm-svn: 112493
2010-08-30 17:47:05 +00:00
e3675ef0f3
Two minor fixes to user-defined literals:
...
- Zero-initialize UDLData so that crashes stop
- Stop complaining that we can't emit them (we most certainly can)
llvm-svn: 112467
2010-08-30 09:27:16 +00:00
79eb5469e0
Implement C++0x user-defined string literals.
...
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
2010-08-29 21:26:48 +00:00
Fariborz Jahanian