Summary:
Previously, we treated CXXUuidofExpr as quite a special case: it was the
only kind of expression that could be a canonical template argument, it
could be a constant lvalue base object, and so on. In addition, we
represented the UUID value as a string, whose source form we did not
preserve faithfully, and that we partially parsed in multiple different
places.

With this patch, we create an MSGuidDecl object to represent the
implicit object of type 'struct _GUID' created by a UuidAttr. Each
UuidAttr holds a pointer to its 'struct _GUID' and its original
(as-written) UUID string. A non-value-dependent CXXUuidofExpr behaves
like a DeclRefExpr denoting that MSGuidDecl object. We cache an APValue
representation of the GUID on the MSGuidDecl and use it from constant
evaluation where needed.

This allows removing a lot of the special-case logic to handle these
expressions. Unfortunately, many parts of Clang assume there are only
a couple of interesting kinds of ValueDecl, so the total amount of
special-case logic is not really reduced very much.

This fixes a few bugs and issues:
 * PR38490: we now support reading from GUID objects returned from
   __uuidof during constant evaluation.
 * Our Itanium mangling for a non-instantiation-dependent template
   argument involving __uuidof no longer depends on which CXXUuidofExpr
   template argument we happened to see first.
 * We now predeclare ::_GUID, and permit use of __uuidof without
   any header inclusion, better matching MSVC's behavior. We do not
   predefine ::__s_GUID, though; that seems like a step too far.
 * Our IR representation for GUID constants now uses the correct IR type
   wherever possible. We will still fall back to using the
      {i32, i16, i16, [8 x i8]}
   layout if a definition of struct _GUID is not available. This is not
   ideal: in principle the two layouts could have different padding.

Reviewers: rnk, jdoerfert

Subscribers: arphaman, cfe-commits, aeubanks

Tags: #clang

Differential Revision: https://reviews.llvm.org/D78171

This moves v32i16/v64i8 to a model consistent with how we
treat integer types with avx1.

This does change the ABI for types vXi16/vXi8 vectors larger than
512 bits to pass in multiple zmms instead of multiple ymms. We'd
already hacked some code to make v64i8/v32i16 pass in zmm.

Cost model is still a bit of a mess. In some place I tried to
match existing behavior. But really we need to account for
splitting and concating costs. Cost model for shuffles is
especially pessimistic.

Differential Revision: https://reviews.llvm.org/D76212

When running the tests through `lit` directly instead of through `check-cxx`,
it is required to manually build the `cxx` (and often `cxx_experimental`)
targets. Instead of having to do that manually, this commit adds a new
target `check-cxx-deps` that does that for you.

This revision introduces a utility to unswitch affine.for/parallel loops
by hoisting affine.if operations past surrounding affine.for/parallel.
The hoisting works for both perfect/imperfect nests and in the presence
of else blocks. The hoisting is currently to as outermost a level as
possible.  Uses a test pass to test the utility.
Add convenience method Operation::getParentWithTrait<Trait>.

Depends on D77487.

Differential Revision: https://reviews.llvm.org/D77870

Since we are going to touch so many lines, do some other cleanups.
Delete Inputs/map-file3.s and Inputs/map-file4.s which are tiny.

Addresses post-commit review feedback from https://reviews.llvm.org/D70265

clang/unittests/AST/ASTImporterFixtures.h:70:5: warning: suggest braces around initialization of subobject [-Wmissing-braces]
    ArgVector(), ArgVector{"-fdelayed-template-parsing"},
    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    {

Ideally types without a corresponding register class wouldn't reach
here, but we're currently missing some (in particular a 192-bit class
is missing).

Baseline for any change due to D75567

Summary:
Since a full run of test-release.sh takes many hours (at least on my
poor systems), we might as well spend some extra time compressing the
tarball, in return for a quite a bit of gains for uploading and
downloading it.

As an example, the 10.0.0-rc4 .tar.xz tarball shrinks from 465MiB to
306MiB, about 52% smaller.

Reviewers: hans, tstellar, rovka

Reviewed By: hans

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D76192

Summary:
Some of the regression tests, such as those for the various sanitizers,
use huge shadow memory maps (showing up in top as 20 TiB).  If any of
those ever crashes, your test system's disk will be filled up until
everything falls over.  Set the ulimit for core dumps to 0 to prevent
this problem.

Reviewers: hans, tstellar, rovka

Reviewed By: hans

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D76191

When building these tests without modules in the whole test suite,
the __config_site macro definitions are not included anymore in the
%{compile_flags}. This causes the _LIBCPP_HAS_NO_THREADS define not
to be picked up, and the test to XPASS on single-threaded systems.

This is a stop-gap measure to fix the build bots, however the proper
solution would be to always pass the __config_site defines as compiler
macros, whether we build with modules or not.

https://reviews.llvm.org/D76210#inline-715185

There's been some back and forth if the cfg paths in the
config_map should be normcase()d. The argument for is that
it allows using all-lower spelling in cmd on Windows, the
argument against that doing so is lossy.

Before the relative-paths-in-generated-lit.site.cfg.py work,
there was no downside to calling normcase(), but with it
we need a hack to recover the original case.

This time, normcase() the hashtable key, but store the original
cased key in addition to the value. This fixes both cons, at the
cost of a few bytes more memory.

Differential Revision: https://reviews.llvm.org/D78169

Summary:
Do line/col to byte conversions on the python side rather than relying on vim.
Its calculations are off when text annotations are present:
- https://github.com/vim/vim/issues/5930
- https://github.com/vim/vim/issues/3718 (fixed, but vim 8.1 is still common)

Reviewers: hokein

Subscribers: cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D78198

There are some inline builtin definitions that we can't emit
(isTriviallyRecursive & callers go into why). Marking these
nobuiltin is only useful if we actually emit the body, so don't mark
these as such unless we _do_ plan on emitting that.

This suboptimality was encountered in Linux (see some discussion on
D71082, and https://github.com/ClangBuiltLinux/linux/issues/979).

Differential Revision: https://reviews.llvm.org/D78162

I believe this call is unneeded after https://reviews.llvm.org/D34855.

Reviewed as part of https://reviews.llvm.org/D78169

MCExpr has a bunch of free space that is currently going to waste.
Repurpose it as 24 bits of subclass data, which is enough to reduce
the size of all subclasses by 8 bytes. This gives us some respectable
savings for debuginfo builds. Here are the max-rss reductions for the
fat LTO link step:

    kc.link               238MiB   231MiB   (-2.82%)
    sqlite3.link          258MiB   250MiB   (-3.27%)
    consumer-typeset.link 152MiB   148MiB   (-2.51%)
    bullet.link           197MiB   192MiB   (-2.30%)
    tramp3d-v4.link       578MiB   567MiB   (-1.92%)
    pairlocalalign.link    92MiB    90MiB   (-1.98%)
    clamscan.link         230MiB   223MiB   (-2.81%)
    lencod.link           242MiB   235MiB   (-2.67%)
    SPASS.link            235MiB   230MiB   (-2.23%)
    7zip-benchmark.link   450MiB   435MiB   (-3.25%)

Differential Revision: https://reviews.llvm.org/D77939

-Consistently name the functions as split*
-Add a helper for doing the two extractSubvector calls and determining the size of the split
-Use getSplitDestVTs to get the result type for the split node.
-Move the binary and unary helper to one place in the file near the extractSubvector functions. Left the VSETCC one near LowerVSETCC since that's its only caller.
-Remove the 256/512 wrappers that just had asserts. I don't think they provided a lot of value and now with the routines called split* the call sites are more obvious what they do.
-Make the unary routine support different source and dest types to support D76212.
-Add some weaker asserts into the helpers to make up for losing the very specific asserts from the 256/512 wrappers.

Differential Revision: https://reviews.llvm.org/D78176

Script to replay reproducers in parallel using the command line driver.
This is used for active replay (stage 1 as described in the RFC on
lldb-dev [1]).

[1] http://lists.llvm.org/pipermail/lldb-dev/2020-April/016100.html

Differential revision: https://reviews.llvm.org/D77608

See also r267736.

We generally only combine starting from users to defs in the artifact combiner,
but this doesn't catch cases where at the point of combining a G_UNMERGE we don't
yet have the opposite G_MERGE on input yet since we haven't legalized that far.

This change adds the users of a G_MERGE to the artifact combiner worklist if one
of the uses is a G_UNMERGE or G_TRUNC.

Differential Revision: https://reviews.llvm.org/D77931

Reviewers: mtrofin, dblaikie

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D78182

Instead of using .sh.cpp tests for the modules tests, use .compile.pass.cpp
and add the -fmodules additional flag.

This reverts commit bddac41b.

Avoid using <sys/types.h> in those tests so that we can run them on
non-AIX systems (otherwise this test is basically dead-code on all
the build bots I'm aware of). Also, split up the test to allow using
.compile.pass.cpp tests instead of .sh.cpp tests, since that is the
last test referencing the %{compile} substitution explicitly.

Summary:
ModelBuilder was missing an api to easily generate yield-for-loops.
This diffs implements an interface allowing to write:
```
%2:2 = loop.for %i = %start to %end step %step iter_args(%arg0 = %init0, %arg1 = %init1) -> (f32, f32) {
  %sum = addf %arg0, %arg1 : f32
  loop.yield %arg1, %sum : f32, f32
}
%3 = addf %2#0, %2#1 : f32
```

as

```
auto results =
    LoopNestBuilder(&i, start, end, step, {&arg0, &arg1},  {init0, init1})([&] {
      auto sum = arg0 + arg1;
      loop_yield(ArrayRef<ValueHandle>{arg1, sum});
    });

// Add the two values accumulated by the yield-for-loop:
ValueHandle(results[0]) + ValueHandle(results[1]);
```

Differential Revision: https://reviews.llvm.org/D78093

As a fly-by fix, improve/fix the comments explaining the tests. This is
a step towards removing references to the %{compile} substitution.

Instead of being ShTests that use clang-verify (and without the proper
REQUIRES annotation), create .verify.cpp tests instead with the right
REQUIRES annotation.

Summary:
This patch adds support for importing fixed point literals, following
up to https://reviews.llvm.org/D46915 specifically for importing AST.

Reviewers: martong, leonardchan, ebevhan, a.sidorin, shafik

Reviewed By: martong

Subscribers: balazske, rnkovacs, teemperor, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D77721

Summary:
The combine for unmerge(cast(merge)) is only valid for vectors, but was
missing a corresponding check. Add a check that the operands are vectors
to avoid an invalid combine.

Without this check, the combiner would emit incorrect code for scalars
and pointers because the artifact cast (trunc/ext) only affects bits at
the end of the type, while this combine assumes that the casted bits
appear between meaningful bits.

This also uncovered a segmentation fault in the AMDGPU
InstructionSelector. The tests triggering this bug have been moved to
their own file and a check for the segmentation fault has been added.

Reviewers: arsenm, dsanders, aemerson, paquette, aditya_nandakumar

Reviewed By: arsenm

Subscribers: tpr, jvesely, wdng, nhaehnle, rovka, kerbowa, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D78191

Reviewed By: broadwaylamb, russell.gallop

Differential Revision: https://reviews.llvm.org/D78153

GCC collect2 passes several options to the linker even if LTO is not used
(note, lld does not support GCC LTO). The lto-wrapper may be a relative
path (especially during development, when gcc is in a build directory), e.g.

  -plugin-opt=relative/path/to/lto-wrapper

We need to ignore such options, which are currently interpreted by
cl::ParseCommandLineOptions() and will fail with `error: --plugin-opt: ld.lld: Unknown command line argument 'relative/path/to/lto-wrapper'`
because the path is apparently not an option registered by an `llvm::cl::opt`.

See lto-plugin-ignore.s for how we interpret various -plugin-opt= options now.

Reviewed By: grimar, tejohnson

Differential Revision: https://reviews.llvm.org/D78158

By renaming .fail.cpp tests that don't need clang-verify to .compile.fail.cpp,
the new test format will not try to compile these tests with clang-verify,
and the old test format will work just the same. However, this allows
removing a workaround that requires parsing each test looking for
clang-verify markup.

After this change, a .fail.cpp test should always have clang-verify markup.
When clang-verify is not supported by the compiler, we will just check that
these tests fail to compile. When clang-verify is supported, these tests
will be compiled with clang-verify whether they have markup or not (so
they should have markup, or they will fail).

This simplifies the test suite and also ensures that all of our .fail.cpp
tests provide clang-verify markup. If it's impossible for a test to have
clang-verify markup, it can be moved to a .compile.fail.cpp test, which
are unconditionally just checked for compilation failure.

Summary:
As a follow up to https://reviews.llvm.org/D29014, add translation
support for freeze.

Introduce a new generic instruction G_FREEZE and translate freeze to it.

Reviewers: dsanders, aqjune, arsenm, aditya_nandakumar, t.p.northover, lebedev.ri, paquette, aemerson

Reviewed By: aqjune, arsenm

Subscribers: fhahn, lebedev.ri, wdng, rovka, hiraditya, jfb, volkan, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D77795

Summary: Update `latest_release` to reflect the release of Clang 10.

Reviewed By: rsmith

Differential Revision: https://reviews.llvm.org/D78172

Currently the alloca does not escape and all stores and the memset can
be removed. Adding a use of the alloca ensures not all stores are
eliminated.