This patch extends the matrix spec to allow matrix-by-scalar division.
Originally support for `/` was left out to avoid ambiguity for the
matrix-matrix version of `/`, which could either be elementwise or
specified as matrix multiplication M1 * (1/M2).
For the matrix-scalar version, no ambiguity exists; `*` is also
an elementwise operation in that case. Matrix-by-scalar division
is commonly supported by systems including Matlab, Mathematica
or NumPy.
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D97857
When using -fprofile-list to selectively apply instrumentation only
to certain files or functions, we may end up with a binary that doesn't
have any counters in the case where no files were selected. However,
because on Linux and Fuchsia, we pass -u__llvm_profile_runtime, the
runtime would still be pulled in and incur some non-trivial overhead,
especially in the case when the continuous or runtime counter relocation
mode is being used. A better way would be to pull in the profile runtime
only when needed by declaring the __llvm_profile_runtime symbol in the
translation unit only when needed.
This approach was already used prior to 9a041a7522, but we changed it
to always generate the __llvm_profile_runtime due to a TAPI limitation.
Since TAPI is only used on Mach-O platforms, we could use the early
emission of __llvm_profile_runtime there, and on other platforms we
could change back to the earlier approach where the symbol is generated
later only when needed. We can stop passing -u__llvm_profile_runtime to
the linker on Linux and Fuchsia since the generated undefined symbol in
each translation unit that needed it serves the same purpose.
Differential Revision: https://reviews.llvm.org/D98061
If --gcc-toolchain is specified, we should detect GCC installation there, and suppress other directories for detection.
Reviewed By: mgorny, manojgupta
Differential Revision: https://reviews.llvm.org/D97894
Summary:
The changes introduced in D87946 changed the API for libomptarget
functions. `__kmpc_push_target_tripcount` was a function in Clang 11.x
but was not given a backward-compatible interface. This change will
require people using Clang 13.x or 12.x to recompile their offloading
programs.
Reviewed By: jdoerfert cchen
Differential Revision: https://reviews.llvm.org/D98358
Add documentation that explains how to extend clang with the new
extensions/features. The guidelines also detail clang's position
about the extension pragmas for the new functionality.
Differential Revision: https://reviews.llvm.org/D97072
This moves code that sets the architecture name
and Float ABI into two new functions in
ToolChains/Arch/ARM.cpp. Greatly simplifying ComputeLLVMTriple.
Some light refactoring in setArchNameInTriple to
move local variables closer to their first use.
Reviewed By: ostannard
Differential Revision: https://reviews.llvm.org/D98253
These are incompatible with opaque pointers. This is in preparation
of dropping this API on the IRBuilder side as well.
Instead explicitly pass the loaded type.
Commit 1b04bdc2f3 added support for
capturing the 'this' pointer in a SEH context (__finally or __except),
But the case in which the 'this' pointer is part of a lambda capture
was not handled properly
Differential Revision: https://reviews.llvm.org/D97687
This errors, but doesn't give source location. We'd need to pass
the Record through several layers to get to the location.
Reviewed By: jrtc27
Differential Revision: https://reviews.llvm.org/D98379
Demonstrate how to generate vadd/vfadd intrinsic functions
1. add -gen-riscv-vector-builtins for clang builtins.
2. add -gen-riscv-vector-builtin-codegen for clang codegen.
3. add -gen-riscv-vector-header for riscv_vector.h. It also generates
ifdef directives with extension checking, base on D94403.
4. add -gen-riscv-vector-generic-header for riscv_vector_generic.h.
Generate overloading version Header for generic api.
https://github.com/riscv/rvv-intrinsic-doc/blob/master/rvv-intrinsic-rfc.md#c11-generic-interface
5. update tblgen doc for riscv related options.
riscv_vector.td also defines some unused type transformers for vadd,
because I think it could demonstrate how tranfer type work and we need
them for the whole intrinsic functions implementation in the future.
Authored-by: Roger Ferrer Ibanez <rofirrim@gmail.com>
Co-Authored-by: Zakk Chen <zakk.chen@sifive.com>
Reviewed By: jrtc27, craig.topper, HsiangKai, Jim, Paul-C-Anagnostopoulos
Differential Revision: https://reviews.llvm.org/D95016
Initially, this flag was meant to only be used through cc1 and not directly
through the clang driver. However, we accidentally ended up using this flag
as a driver flag already for selecting multilibs within the fuchsia toolchain.
We're currently in an awkward state where it's only accepted as a driver flag
when targeting Fuchsia, and all other instances it can only be added via
-Xclang. Since we're ready to use this in Fuchsia, we can just expose this to
the driver for simplicity.
Differential Revision: https://reviews.llvm.org/D98375
Updates __is_unsigned to have the same behavior as the standard
specifies. This is in line with 511dbd8, which applied the same change
to __is_signed.
Refs D67897.
Differential Revision: https://reviews.llvm.org/D98104
Generate a json file containing descriptions of AST classes and their
public accessors which return SourceLocation or SourceRange.
Use the JSON file to generate a C++ API and implementation for accessing
the source locations and method names for accessing them for a given AST
node.
This new API can be used to implement 'srcloc' output in clang-query:
http://ce.steveire.com/z/m_kTIo
In this first version of this feature, only the accessors for Stmt
classes are generated, not Decls, TypeLocs etc. Those can be added
after this change is reviewed, as this change is mostly about
infrastructure of these code generators.
Differential Revision: https://reviews.llvm.org/D93164
The patch adds an argument to update_cc_test_checks for replacing a function name matching a regex. This functionality is needed to match generated function signatures that include file hashes. Example:
The function signature for the following function:
`__omp_offloading_50_b84c41e__Z9ftemplateIiET_i_l30_worker`
with `--replace-function-regex "__omp_offloading_[0-9]+_[a-z0-9]+_(.*)"` will become:
`CHECK-LABEL: @{{__omp_offloading_[0-9]+_[a-z0-9]+__Z9ftemplateIiET_i_l30_worker}}(`
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D97107
Some tests in clang require running non-filechecked commands to generate the actual filecheck input. For example, tests for openmp offloading require generating the host bc without any checking, before running the clang command to actually generate the filechecked IR of the target device. This patch enables `update_cc_test_checks.py` to run non-filechecked run lines in-place.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D97068
When building with `LLVM_BUILD_EXTERNAL_COMPILER_RT=ON` (e.g. Swift does
this) we do an "external" build of compiler-rt where we build
compiler-rt with the just built clang.
Unfortunately building in this mode had a bug where compiler-rt would
not get rebuilt if compiler-rt sources changed. This is problematic
for incremental builds because it meant that the compiler-rt binaries
were stale.
The fix is to use the `BUILD_ALWAYS` ExternalProject_Add option which
means the build command for compiler-rt is always run.
In principle if all of the following are true:
* compiler-rt has already been built.
* there are no compiler-rt source changes.
* the compiler hasn't changed.
* ninja is being used as the generator for the compiler-rt build.
then the overhead for always running the build command for incremental
builds is negligible.
However, in practice clang gets rebuilt everytime the HEAD commit
changes (due to commit hash being embedded in the output of `--version`)
which means all of compiler-rt will be rebuilt everytime this happens.
While this is annoying it's better to do the slow but correct thing
rather than the fast but incorrect thing.
rdar://75150660
Differential Revision: https://reviews.llvm.org/D98291
This patch adds `-fdebug-dump-parsing-log` in the new driver. This option is
semantically identical to `-fdebug-instrumented-parse` in `f18` (the
former is added as an alias in `f18`).
As dumping the parsing log makes only sense for instrumented parses, we
set Fortran::parser::Options::instrumentedParse to `True` when
`-fdebug-dump-parsing-log` is used. This is consistent with `f18`.
To facilitate tweaking the configuration of the frontend based on the
action being requested, `setUpFrontendBasedOnAction` is introduced in
CompilerInvocation.cpp.
Differential Revision: https://reviews.llvm.org/D97457
If the non-iterator side of an iterator operation
`+`, `+=`, `-` or `-=` is `UndefinedVal` an assertions happens.
This small fix prevents this.
Patch by Adam Balogh.
Reviewed By: NoQ
Differential Revision: https://reviews.llvm.org/D85424
Before `bc713f6a004723d1325bc16e1efc32d0ac82f939` landed, the analyzer
crashed on this reduced example.
It seems important to have bot `ctu` and `-analyzer-opt-analyze-headers`
enabled in the example.
This test file ensures that no regression happens in the future in this regard.
Reviewed By: martong, NoQ
Differential Revision: https://reviews.llvm.org/D96586
According to a Bugzilla ticket (https://bugs.llvm.org/show_bug.cgi?id=45148),
ArrayBoundCheckerV2 produces a false-positive report.
This patch adds a test demonstrating the current //flawed// behavior.
Also adds several similar test cases just to be on the safe side.
Reviewed By: martong
Differential Revision: https://reviews.llvm.org/D86870
IR produced using TableGen builtin function declarations
(`fdeclare-opencl-builtins.cl`) did not have the target's calling
convention applied to builtin calls.
Fix this, and update the codegen test to check that IR produced using
opencl-c.h and `-fdeclare-opencl-builtins` is identical with respect
to the builtin calls.
Differential Revision: https://reviews.llvm.org/D98039
It's available both in CodeGenOptions and in LangOptions, and LangOptions
implementation is slightly better as it uses a StringRef instead of a char
pointer, so use it.
Differential Revision: https://reviews.llvm.org/D98175
`initFunctionSummaries` lazily initializes a data structure with
function summaries for standard library functions. It is called for
every pre-, post-, and eval-call events, i.e. 3 times for each call on
the path. If the initialization doesn't find any standard library
functions in the translation unit, it will get re-tried (with the same
effect) many times even for small translation units.
For projects not using standard libraries, the speed-up can reach 50%
after this patch.
Differential Revision: https://reviews.llvm.org/D98244
In -fno-exceptions -fno-asynchronous-unwind-tables -g0 mode,
GCC does not emit `.cfi_*` directives.
```
% diff <(gcc -fno-asynchronous-unwind-tables -dM -E a.c) <(gcc -dM -E a.c)
130a131
> #define __GCC_HAVE_DWARF2_CFI_ASM 1
```
This macro is useful because code can decide whether inline asm should include `.cfi_*` directives.
`.cfi_*` directives without `.cfi_startproc` can cause assembler errors
(integrated assembler: `this directive must appear between .cfi_startproc and .cfi_endproc directives`).
Differential Revision: https://reviews.llvm.org/D97743
GCC warning:
```
/llvm-project/clang/lib/APINotes/APINotesYAMLCompiler.cpp:574:6: warning: ‘void {anonymous}::Module::dump()’ defined but not used [-Wunused-function]
574 | void Module::dump() {
| ^~~~~~
```
By default, the driver uses the compiler-rt builtins and links with
-l:libunwind.a.
Restore the previous behavior by passing --rtlib=libgcc.
Reviewed By: danalbert
Differential Revision: https://reviews.llvm.org/D96404
The only time we would consider allowing this is inside a call to
std::allocator<T>::deallocate, whose contract does not permit deletion
of null pointers.
These functions were local to SemaDeclAttr.cpp, but these functions are
useful in general (for instance, for statement or type attribute
processing). This refactoring is in advance of beginning to tablegen
diagnostic checks for statement attributes the way we already do for
declaration attributes.
There is one functional change in here as a drive-by. The
external_source_symbol attribute had one of its diagnostic checks
inside of an assert, which was corrected.
In -fno-exceptions -fno-asynchronous-unwind-tables -g0 mode,
GCC does not emit `.cfi_*` directives.
```
% diff <(gcc -fno-asynchronous-unwind-tables -dM -E a.c) <(gcc -dM -E a.c)
130a131
> #define __GCC_HAVE_DWARF2_CFI_ASM 1
```
This macro is useful because code can decide whether inline asm should include `.cfi_*` directives.
`.cfi_*` directives without `.cfi_startproc` can cause assembler errors
(integrated assembler: `this directive must appear between .cfi_startproc and .cfi_endproc directives`).
Differential Revision: https://reviews.llvm.org/D97743
LLVM is recommending to use SmallVector (that is, omitting the N), in the
absence of a well-motivated choice for the number of inlined elements N.
However, this doesn't work well with XL compiler on AIX since some header(s)
aren't properly picked up with it. We need to take a further look into the real
issue underneath and fix it in a later patch.
But currently we'd like to use this patch to unblock the build compiler issue
first.
Differential Revision: https://reviews.llvm.org/D98265
We used to trigger assertion when transforming c-tor with unparsed
default argument. Now we ignore such constructors for this purpose.
Differential Revision: https://reviews.llvm.org/D97965
In case a char-literal of type int (C/ObjectiveC) corresponds to a
format specifier with the %hh length modifier, don't treat the literal
as of type char for issuing diagnostics, as otherwise this results in:
printf("%hhd", 'e');
warning: format specifies type 'char' but the argument has type 'char'.
Differential revision: https://reviews.llvm.org/D97951
SUMMARY:
n the patch https://reviews.llvm.org/D87451 "add new option -mignore-xcoff-visibility"
we did as "The option -mignore-xcoff-visibility has no effect on visibility attribute when compile with -emit-llvm option to generated LLVM IR."
in these patch we let -mignore-xcoff-visibility effect on generating IR too. the new feature only work on AIX OS
Reviewer: Jason Liu,
Differential Revision: https://reviews.llvm.org/D89986
This patch implements the conditional select operator for
ext_vector_types in C++. It does so by using the same semantics as for
C.
D71463 added support for the conditional select operator for VectorType
in C++. Unfortunately the semantics between ext_vector_type in C are
different to VectorType in C++. Select for ext_vector_type is based on
the MSB of the condition vector, whereas for VectorType it is `!= 0`.
This unfortunately means that the behavior is inconsistent between
ExtVectorType and VectorType, but I think using the C semantics for
ExtVectorType in C++ as well should be less surprising for users.
Reviewed By: erichkeane, aaron.ballman
Differential Revision: https://reviews.llvm.org/D98055
Builtins that require multiple extensions, such as certain
`write_imagef` forms, were not exposed because of the Sema check not
splitting the extension string.
Differential Revision: https://reviews.llvm.org/D97930
See https://bugs.llvm.org/show_bug.cgi?id=42154.
GCC's __attribute__((align)) can reduce the alignment of a type when applied to
a typedef. However, functions which take a pointer or reference to the
original type are compiled assuming the original alignment. Therefore when any
such function is passed an object of the new, less-aligned type, an alignment
fault can occur. In particular, this applies to the constructor, which is
defined for the original type and called for the less-aligned object.
This change adds a warning whenever an pointer or reference to an object is
passed to a function that was defined for a more-aligned type.
The calls to ASTContext::getTypeAlignInChars seem change the order in which
record layouts are evaluated, which caused changes to the output of
-fdump-record-layouts. As such some tests needed to be updated:
* Use CHECK-LABEL rather than counting the number of "Dumping AST Record
Layout" headers.
* Check for end of line in labels, so that struct B1 doesn't match struct B
etc.
* Add --strict-whitespace, since the whitespace shows meaningful structure.
* The order in which record layouts are printed has changed in some cases.
* clang-format for regions changed
Differential Revision: https://reviews.llvm.org/D97187
This updates the canonical text proto raw string delimiter to `pb` for Google style, moving codebases towards a simpler and more consistent style.
Also updates a behavior where the canonical delimiter was not applied for raw strings with empty delimiters detected via well-known enclosing functions that expect a text proto, effectively making the canonical delimiter more viral. This feature is not widely used so this should be safe and more in line with promoting the canonicity of the canonical delimiter.
Reviewed By: sammccall
Differential Revision: https://reviews.llvm.org/D97688
This is the first patch supporting M68k in Clang
- Register M68k as a target
- Target specific CodeGen support
- Target specific attribute support
Authors: myhsu, m4yers, glaubitz
Differential Revision: https://reviews.llvm.org/D88393
002dd47bdd was meant to not be any
functional change, but it turned out it was.
With CLANG_DEFAULT_RTLIB set to compiler-rt, CLANG_DEFAULT_UNWINDLIB used
to bet set to an empty string, but now was set to "none".
If one only overrode rtlib to libgcc, one previously would get libgcc
as unwind lib, but now didn't. This caused test failures, fixed in
41476d89b8.
Secondly, for the android target, the previous default was to link
libunwind, which this now changed.
Reinstate the exact same behaviour as before (removing the previously
typoed cmake check) and fix the option comment in one place to match
the other one above.
Differential Revision: https://reviews.llvm.org/D98142
In D97003, CUDA 9.2 is the minimum requirement for OpenMP offloading on
NVPTX target. We don't need to have macros in source code to select right functions
based on CUDA version. we don't need to compile multiple bitcode libraries of
different CUDA versions for each SM. We don't need to worry about future
compatibility with newer CUDA version.
`-target-feature +ptx61` is used in this patch, which corresponds to the highest
PTX version that CUDA 9.2 can support.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D97198
Some tests in clang require running non-filechecked commands to generate the actual filecheck input. For example, tests for openmp offloading require generating the host bc without any checking, before running the clang command to actually generate the filechecked IR of the target device. This patch enables `update_cc_test_checks.py` to run non-filechecked run lines in-place.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D97068
This changes the target data layout to make stack align to 16 bytes
on Power10. Before this change, stack was being aligned to 32 bytes.
Reviewed By: #powerpc, nemanjai
Differential Revision: https://reviews.llvm.org/D96265
From `man ld`:
-why_load Log why each object file in a static library is loaded.
That is, what symbol was needed.
Also called -whyload for compatibility.
`-why_load` is the spelling preferred by the linker and `-whyload` an old
compatibility setting. clang should accept the preferred form, and map both
forms to the preferred form.
Differential Revision: https://reviews.llvm.org/D98156
Spack is a package management tool extensively used by HPC community.
As ROCm packages are built by Spack by HPC community, we need to teach
clang driver to detect ROCm installation built by Spack.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D97340
For MinGW targets, we distinguish between an explicitly shared unwinder
library (requested via -shared-libgcc), an explicitly static one
(requested via -static-libgcc or -static) and the default case (which
just passes -lunwind to the linker, which will pick either shared or
static depending on what's available, with the normal linker logic).
This makes the implicit default case (as added in D79995) actually work as
it was intended, when using the g++ driver (which is the main usecase for
libunwind as far as I know).
Differential Revision: https://reviews.llvm.org/D98023
CLANG_DEFAULT_RTLIB had a typo, and libunwind isn't a valid
option for it.
This keeps the actual behaviour from before, defaulting to none if
using compiler-rt as rtlib.
Differential Revision: https://reviews.llvm.org/D98022
Prior to this fix, constrained decltype(auto) behaves exactly the same
as constrained regular auto.
This fixes it so it deduces like decltype(auto).
Signed-off-by: Matheus Izvekov <mizvekov@gmail.com>
Reviewed By: rsmith
Differential Revision: https://reviews.llvm.org/D98087
We've observed this test being significantly flaky on our Mac CI
machines when we're running the full check-clang suite. It fails because
the wait_for condition isn't met within 3 seconds. We believe it's
because our CI machines are somewhat underpowered and pretty heavily
loaded when we're running the full check-clang suite.
I ran some experiments on increasing the timeout. I ran the full
check-clang suite 100 times with each timeout value and recorded how
many flaky failures we encountered in these tests. The results are:
3 second timeout (baseline): 20 failures
10 second timeout: 14 failures
20 second timeout: 4 failures
30 second timeout: 2 failures
40 second timeout: 1 failure
50 second timeout: 0 failures
60 second timeout: 0 failures
I ran another set of 100 tests for the 50 second timeout and observed
one flaky failure. By contrast, I ended up running check-clang 500 times
for the 60 second timeout and didn't observe a single flaky failure.
That's how the 60 second timeout value used in this patch was derived.
While a 60 second timeout might seem high, keep in mind that:
- This is a timeout, not a sleep; the test should require much less time
the vast majority of instances, especially on more powerful machines.
- The long timeout is most likely to occur when other tests are also
running at the same time, so the latency of the timeout will also be
masked by the latency of the other tests.
See https://reviews.llvm.org/D58418?id=200123#inline-554211 for where
this timeout was originally introduced and the possibility of raising it
if it wasn't enough was discussed.
Reviewed By: plotfi
Differential Revision: https://reviews.llvm.org/D97878
Fix duplicate diagnostic for an over-aligned allocation with no matching
function, and add custom diagnostic for the case where the
non-allocating placement new was intended but <new> was not included.
The option -funique-internal-linkage-names was added in D73307 and D78243 as a
LLVM early pass to insert a unique suffix to internal linkage functions and
vars. The unique suffix was the hash of the module path. However, we found
that this can be done more cleanly in clang early and the fixes that need to
be done later can be completely avoided. The fixes in particular are trying
to modify the DW_AT_linkage_name and finding the right place to insert the
pass.
This patch ressurects the original implementation proposed in D73307 which
was reviewed and then ditched in favor of the pass based approach.
Differential Revision: https://reviews.llvm.org/D96109
This commit removes the old way of handling Whitesmiths mode in favor of just setting the
levels during parsing and letting the formatter handle it from there. It requires a bit of
special-casing during the parsing, but ends up a bit cleaner than before. It also removes
some of switch/case unit tests that don't really make much sense when dealing with
Whitesmiths.
Differential Revision: https://reviews.llvm.org/D94500
This amends 25f753c51e.
When applying the child configurations we don't need any diagnostic,
because it was issued when first parsing them. So just drop everything
on the second parse.
Differential Revision: https://reviews.llvm.org/D96760
emission
Ensure that we are in a function declaration context before checking
the diagnostic emission status, to avoid dereferencing a NULL function
declaration.
Differential Revision: https://reviews.llvm.org/D97573
gfx1030 added a new way to implement readcyclecounter using the
SHADER_CYCLES hardware register, but the s_memtime instruction still
exists, so the MC layer should still accept it and the
llvm.amdgcn.s.memtime intrinsic should still work.
Differential Revision: https://reviews.llvm.org/D97928
A bug was introduced when adding -munsafe-fp-atomics.
By default it should be off.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D97967
Like nvptx and some other targets, -mconstructor-aliases does not work well with amdgpu,
therefore we disable it in the same approach.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D97959
`mix` is subtly different from `clamp`: in the overloads where the
last argument is a scalar, the second argument should be a gentype for
`mix`.
As scalars can be implicitly converted to vectors, this cannot be
caught in the Sema test. Hence adding a CodeGen test, where we can
verify the types using the mangled name.
Previously if we couldn't run the clang-format command
for some reason, you'd get an unhelpful error message:
```
OSError: [Errno 2] No such file or directory
```
Which doesn't tell you what was happening to cause this.
Catch the error and add the command we were attempting to run:
```
RuntimeError: Failed to run "<...>/clang-food <...>" - No such file or directory"
RuntimeError: Failed to run "<...>/clang-format <...>" - Permission denied"
```
Reviewed By: krasimir
Differential Revision: https://reviews.llvm.org/D98032
Clang exposes an interface for extending the PCM/PCH file format: `ModuleFileExtension`.
Clang itself has only a single implementation of the interface: `TestModuleFileExtension` that can be instantiated via the `-ftest-module-file_extension=` command line argument (and is stored in `FrontendOptions::ModuleFileExtensions`).
Clients of the Clang library can extend the PCM/PCH file format by pushing an instance of their extension class to the `FrontendOptions::ModuleFileExtensions` vector.
When generating the `-ftest-module-file_extension=` command line argument from `FrontendOptions`, a downcast is used to distinguish between the Clang's testing extension and other (client) extensions.
This functionality is enabled by LLVM-style RTTI. However, this style of RTTI is hard to extend, as it requires patching Clang (adding new case to the `ModuleFileExtensionKind` enum).
This patch switches to the LLVM RTTI for open class hierarchies, which allows libClang users (e.g. Swift) to create implementations of `ModuleFileExtension` without patching Clang. (Documentation of the feature: https://llvm.org/docs/HowToSetUpLLVMStyleRTTI.html#rtti-for-open-class-hierarchies)
Reviewed By: artemcm
Differential Revision: https://reviews.llvm.org/D97702
This patch fixes failure of the `CodeGen/aix-ignore-xcoff-visibility.cpp` test with command line round-trip.
The absence of '-fvisibility' implies '-mignore-xcoff-visibility'.
The problem is that when '-fvisibility default' is passed to -cc1, it isn't being generated. (This adheres to the principle that generation doesn't produce arguments with default values.)
However, that caused '-mignore-xcoff-visibility' to be implied in the generated command line (without '-fvisibility'), while it wasn't implied in the original command line (with '-fvisibility').
This patch fixes that by always generating '-fvisibility' and explains the situation in comment.
(The '-mginore-xcoff-visibility' option was added in D87451).
Reviewed By: Bigcheese
Differential Revision: https://reviews.llvm.org/D97552
David Blaikie