Handle sse turning on mmx, but no -mmx not turning off SSE.
Rationale :
// sse3
__m128d test_mm_addsub_pd(__m128d A, __m128d B) {
return _mm_addsub_pd(A, B);
}
// mmx
void shift(__m64 a, __m64 b, int c) {
_mm_slli_pi16(a, c);
_mm_slli_pi32(a, c);
_mm_slli_si64(a, c);
_mm_srli_pi16(a, c);
_mm_srli_pi32(a, c);
_mm_srli_si64(a, c);
_mm_srai_pi16(a, c);
_mm_srai_pi32(a, c);
}
clang -msse3 -mno-mmx file.c -c
For this code we should be able to explicitly turn off MMX
without affecting the compilation of the SSE3 function and then
diagnose and error on compiling the MMX function.
This is a preparatory patch to the actual diagnosis code which is
coming in a future patch. This sets us up to have the correct information
where we need it and verifies that it's being emitted for the backend
to handle.
Migrate most feature map inclusion to initFeatureMap for the x86 target so
that we can build up an accurate set of features rather than relying on
TargetInfo initialization via handleTargetFeatures to munge the list
of features.
Aaron Ballman [Thu, 8 Oct 2015 19:24:08 +0000 (19:24 +0000)]
When mapping no_sanitize_* attributes to no_sanitize attributes, handle GNU-style formatting that involves prefix and suffix underscores. Cleans up other usages of similar functionality.
Reid Kleckner [Thu, 8 Oct 2015 17:29:07 +0000 (17:29 +0000)]
[clang-cl] Make /EHs turn on C++ EH once again
C++ exceptions are still off by default, which is similar to how C++
cleanups are off by default in MSVC.
If you use clang instead of clang-cl, exceptions are also still off by
default. In the future, when C++ EH is proven to be stable, we may flip
the default for that driver to be consistent with other platforms.
Renato Golin [Thu, 8 Oct 2015 16:43:26 +0000 (16:43 +0000)]
Simplify DefaultCPU in ARMTargetInfo
Simplifying the convoluted CPU handling in ARMTargetInfo.
The default base CPU on ARM is ARM7TDMI, arch ARMv4T, and
ARMTargetInfo had a different one. This wasn't visible from
Clang because the driver selects the defaults and sets the
Arch/CPU features directly, but the constructor depended
on the CPU, which was never used.
This patch corrects the mistake and greatly simplifies
how CPU is dealt with (essentially by removing the duplicated
DefaultCPU field).
David Majnemer [Thu, 8 Oct 2015 10:04:46 +0000 (10:04 +0000)]
[MSVC Compat] Try to treat an implicit, fixed enum as an unfixed enum
consider the following:
enum E *p;
enum E { e };
The above snippet is not ANSI C because 'enum E' has not bee defined
when we are processing the declaration of 'p'; however, it is a popular
extension to make the above work. This would fail using the Microsoft
enum semantics because the definition of 'E' would implicitly have a
fixed underlying type of 'int' which would trigger diagnostic messages
about a mismatch between the declaration and the definition.
Instead, treat fixed underlying types as not fixed for the purposes of
the diagnostic.
Alexey Bataev [Thu, 8 Oct 2015 09:10:53 +0000 (09:10 +0000)]
[OPENMP 4.1] Codegen for array sections/subscripts in 'reduction' clause.
OpenMP 4.1 adds support for array sections/subscripts in 'reduction' clause. Patch adds codegen for this feature.
David Majnemer [Thu, 8 Oct 2015 07:45:35 +0000 (07:45 +0000)]
[Sema] Tweak incomplete enum types on MSVC ABI targets
Enums without an explicit, fixed, underlying type are implicitly given a
fixed 'int' type for ABI compatibility with MSVC. However, we can
enforce the standard-mandated rules on these types as-if we didn't know
this fact if the tag is not part of a definition.
David Majnemer [Thu, 8 Oct 2015 06:31:22 +0000 (06:31 +0000)]
Update tests touched by r249656
These test updates almost exclusively around the change in behavior
around enum: enums without a definition are considered incomplete except
when targeting MSVC ABIs. Since these tests are interested in the
'incomplete-enum' behavior, restrict them to %itanium_abi_triple.
David Majnemer [Thu, 8 Oct 2015 04:53:31 +0000 (04:53 +0000)]
[MSVC Compat] Enable ABI impacting non-conforming behavior independently of -fms-compatibility
No ABI for C++ currently makes it possible to implement the standard
100% perfectly. We wrongly hid some of our compatible behavior behind
-fms-compatibility instead of tying it to the compiler ABI.
Douglas Katzman [Thu, 8 Oct 2015 04:24:12 +0000 (04:24 +0000)]
Stop messing with the 'g' group of options in CompilerInvocation.
With this change, most 'g' options are rejected by CompilerInvocation.
They remain only as Driver options. The new way to request debug info
from cc1 is with "-debug-info-kind={line-tables-only|limited|standalone}"
and "-dwarf-version={2|3|4}". In the absence of a command-line option
to specify Dwarf version, the Toolchain decides it, rather than placing
Toolchain-specific logic in CompilerInvocation.
Also fix a bug in the Windows compatibility argument parsing
in which the "rightmost argument wins" principle failed.
Richard Smith [Thu, 8 Oct 2015 00:17:59 +0000 (00:17 +0000)]
When pretty-printing a C++11 literal operator, don't insert whitespace between
the "" and the suffix; that breaks names such as 'operator""if'. For symmetry,
also remove the space between the 'operator' and the '""'.
Reid Kleckner [Thu, 8 Oct 2015 00:17:45 +0000 (00:17 +0000)]
Don't emit exceptional stackrestore cleanups around inalloca functions
The backend restores the stack pointer after recovering from an
exception. This is similar to r245879, but it doesn't try to use the
normal cleanup mechanism, so hopefully it won't cause the same breakage.
Ehsan Akhgari [Thu, 8 Oct 2015 00:01:20 +0000 (00:01 +0000)]
Make clang_Cursor_getMangling not mangle if the declaration isn't mangled
Right now clang_Cursor_getMangling will attempt to mangle any
declaration, even if the declaration isn't mangled (extern C). This
results in a partially mangled name which isn't useful for much. This
patch makes clang_Cursor_getMangling return an empty string if the
declaration isn't mangled.
Reid Kleckner [Wed, 7 Oct 2015 21:03:41 +0000 (21:03 +0000)]
[WinEH] Don't use lifetime markers for MS catch parameters
We don't have a good place to put them. Our previous spot was causing us
to optimize loads from the exception object to undef, because it was
after the catchpad instruction that models the write to the catch
object.
Daniel Jasper [Wed, 7 Oct 2015 19:56:12 +0000 (19:56 +0000)]
ASTMatchers: Keep AllCallbacks in a set instead of a vector
AllCallbacks is currently only used to call onStartOfTranslationUnit and
onEndOfTranslationUnit on them. In this (and any other scenario I can
come up with), it is important (or at least better) not to have
duplicates in this container. E.g. currently onEndOfTranslationUnit is
called repeatedly on the same callback for every matcher that is
registered with it.
Alexey Bataev [Wed, 7 Oct 2015 10:22:08 +0000 (10:22 +0000)]
Fix crash in codegen on casting to `bool &`.
Currently codegen crashes trying to emit casting to bool &. It happens because bool type is converted to i1 and later then lvalue for reference is converted to i1*. But when codegen tries to load this lvalue it crashes trying to load value from this i1*.
Benjamin Kramer [Wed, 7 Oct 2015 10:05:44 +0000 (10:05 +0000)]
[VFS] Refactor VFSFromYAML a bit.
- Rename it to RedirectingFileSystem. This is what it does, YAML is just a
serialization format for it.
- Consistently use unique_ptr for memory management.
Ehsan Akhgari [Tue, 6 Oct 2015 18:24:33 +0000 (18:24 +0000)]
Make clang_Cursor_getMangling don't mangle if the declaration isn't mangled
Right now clang_Cursor_getMangling will attempt to mangle any
declaration, even if the declaration isn't mangled (extern "C"). This
results in a partially mangled name which isn't useful for much. This
patch makes clang_Cursor_getMangling return an empty string if the
declaration isn't mangled.
Benjamin Kramer [Tue, 6 Oct 2015 14:45:20 +0000 (14:45 +0000)]
[Tooling] Reuse FileManager in ASTUnit.
ASTUnit was creating multiple FileManagers and throwing them away. Reuse
the one from Tooling. No functionality change now but necessary for
VFSifying tooling.
Benjamin Kramer [Tue, 6 Oct 2015 14:45:13 +0000 (14:45 +0000)]
[Tooling] Remove dead code.
It took me some time to figure out why this is not working as expected:
std:error_code converts to true if there is an error. This means we never
ever took the generated absolute path, which happens to be the right thing
anyways as it properly works with virtual files. Just remove the whole
thing, relative paths are covered by existing tooling tests.
Manuel Klimek [Tue, 6 Oct 2015 10:45:03 +0000 (10:45 +0000)]
Adds a way for tools to deduce the target config from a compiler name.
Adds `addTargetAndModeForProgramName`, a utility function that will add
appropriate `-target foo` and `--driver-mode=g++` tokens to a command
line for driver invocations of the form `a/b/foo-g++`. It is intended to
support tooling: for example, should a compilation database record some
invocation of `foo-g++` without these implicit flags, a Clang tool may
use this function to add them back.
Benjamin Kramer [Tue, 6 Oct 2015 10:04:08 +0000 (10:04 +0000)]
[VFS] Transition clang-format to use an in-memory FS.
Apart from being cleaner this also means that clang-format no longer has
access to the host file system. This isn't necessary because clang-format
never reads includes :)
Includes minor tweaks and bugfixes found in the VFS implementation while
running clang-format tests.
Richard Smith [Mon, 5 Oct 2015 20:05:21 +0000 (20:05 +0000)]
Revert r107690 (for PR7417) and add a testcase that it breaks. The approach of
that change turns out to not be reasonable: mutating the AST of a parsed
template during instantiation is not a sound thing to do, does not work across
chained PCH / modules builds, and is in any case a special-case workaround to a
more general problem that should be solved centrally.
Adrian Prantl [Mon, 5 Oct 2015 17:41:16 +0000 (17:41 +0000)]
Undo the unique_ptr'fication of CodeGenABITypes::CGM introduced in r248762.
include/clang/CodeGenABITypes.h is in meant to be included by external
users, but using a unique_ptr on the private CodeGenModule introduces a
dependency on the type definition that prevents such a use.
Benjamin Kramer [Mon, 5 Oct 2015 13:55:20 +0000 (13:55 +0000)]
[VFS] Add working directories to every virtual file system.
For RealFileSystem this is getcwd()/chdir(), the synthetic file systems can
make up one for themselves. OverlayFileSystem now synchronizes the working
directories when a new FS is added to the overlay or the overlay working
directory is set. This allows purely artificial file systems that have zero
ties to the underlying disks.
Benjamin Kramer [Mon, 5 Oct 2015 13:55:14 +0000 (13:55 +0000)]
[VFS] Add an in-memory file system implementation.
This is a simple file system tree of memory buffers that can be filled by a
client. In conjunction with an OverlayFS it can be used to make virtual
files accessible right next to physical files. This can be used as a
replacement for the virtual file handling in FileManager and which I intend
to remove eventually.
Re-commit "Add support for the new mips-mti-linux toolchain."
r249137 added support for the new mips-mti-linux toolchain. However,
the new tests of that commit, broke some buildbots because they didn't use
the correct regular expressions to capture the filename of Clang & LLD.
This commit re-applies the changes of r249137 and fixes the tests in
r249137 in order to match the filenames of the Clang and LLD executable.
Daniel Jasper [Mon, 5 Oct 2015 07:36:40 +0000 (07:36 +0000)]
clang-format: Remove sentence in option cost threshold that weakens the
statement. Specifically, we don't want people that have already written
a patch to just write a style guide for their 1-person project.
Keno Fischer [Sun, 4 Oct 2015 18:51:04 +0000 (18:51 +0000)]
Fix makefile build on OSX when ARM targets are not enabled
Summary: When LLVM/Clang is built without ARM support, the ios_kext runtime
library is not built, but without this patch, the Makefile still tries to
copy it. This is a recent regression, because the ios_kext library used
to also be built on x86_64.
Add -f[no-]declspec to control recognition of __declspec as a keyword
In versions of clang prior to r238238, __declspec was recognized as a keyword in
all modes. It was then changed to only be enabled when Microsoft or Borland
extensions were enabled (and for CUDA, as a temporary measure). There is a
desire to support __declspec in Playstation code, and possibly other
environments. This commit adds a command-line switch to allow explicit
enabling/disabling of the recognition of __declspec as a keyword. Recognition
is enabled by default in Microsoft, Borland, CUDA, and PS4 environments, and
disabled in all other environments.
Eric Fiselier [Sun, 4 Oct 2015 00:11:02 +0000 (00:11 +0000)]
Diagnose const atomics in __atomic builtins.
Diagnose when a pointer to const T is used as the first argument in at atomic
builtin unless that builtin is a load operation. This is already checked for
C11 atomics builtins but not for __atomic ones.
This patch was given the LGTM by rsmith when it was part
of a larger review. (See http://reviews.llvm.org/D10407)
The default target is ARM on the ARM self host bots. This is problematic since
the behaviour on x86, x64 is different from ARM. Explicitly pass the target.
This should hopefully fix the ARM bots.
The Windows on ARM ABI recommends that FPO be disabled. This is since the
Windows on ARM ABI uses the FP for fast stack walking. By paying the slight
cost of the loss of registers, a much faster backtrace is possible by using the
frame pointer since the pdata need not be consulted. Furthermore, even if pdata
is not available, you can still more easily reconstruct the stack.
Ahmed Bougacha [Fri, 2 Oct 2015 22:41:59 +0000 (22:41 +0000)]
Don't nil check non-nil class receiver of AArch64 stret calls.
I randomly came across this difference between AArch64 and other targets:
on the latter, we don't emit nil checks for known non-nil class method
calls thanks to r247350, but we still do for AArch64 stret calls.
They use different code paths, because those are special, as they go
through the regular msgSend, not the msgSend*_stret variants.