Differential Revision: https://reviews.llvm.org/D146740

Vector dialect patterns have grown enormously in the past year to a point where they are now impenetrable.
Start reorganizing them towards finer-grained control.

Differential Revision: https://reviews.llvm.org/D146736

My recent change [1] extended the external-swift-debugging.cpp test, but
didn't account for PAC under which function pointers aren't trivially
comparable. We could use `ptrauth_strip()`, but for the test it's easier
to just the symbol name.

[1] https://reviews.llvm.org/D146264

Differential Revision: https://reviews.llvm.org/D146742

Reviewed By: mkazantsev

Differential Revision: https://reviews.llvm.org/D145705

We don't need an explicit AND mask, we can use KnownBits to determine if each element has (the same) single non-zero bit and shift that into the msb/signbit for MOVMSK to access directly.

When fixing the test earlier, we missed the JSON case for NaN and INF,
so handle those the same as for non-JSON, by creating the string
dynamically.

Reviewed By: abhina.sreeskantharajan

Differential Revision: https://reviews.llvm.org/D146739

The trailing return type arrow checker verifies that a declaration is
being parsed, however, this isn't true when inside of macros.

It turns out the existence of the auto keyword is enough to make
sure that we're dealing with a trailing return type, and whether we're
in a declaration doesn't matter.

Fixes https://github.com/llvm/llvm-project/issues/47664

Reviewed By: HazardyKnusperkeks, owenpan

Differential Revision: https://reviews.llvm.org/D141811

Reviewed By: mkazantsev

Differential Revision: https://reviews.llvm.org/D143892

Building with -DLIBCXX_ENABLE_THREADS=OFF -DLIBCXXABI_ENABLE_THREADS=OFF
(like e.g. for wasm) fails after D146228 because of a misplaced std
namespace begin/end.

Reviewed By: philnik, #libc

Differential Revision: https://reviews.llvm.org/D146682

Implemented [[ https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2022/p2711r1.html | P2711R1 ]] for existing views.
 (`join_with_view` is not yet implemented)

Reviewed By: #libc, philnik

Differential Revision: https://reviews.llvm.org/D144822

Close #61322

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D146603

Differential Revision: https://reviews.llvm.org/D146738

The printf and fprintf implementations use our internal implementation
to improve performance when it's available, but this patch enables using
the public FILE API for overlay mode.

Reviewed By: sivachandra, lntue

Differential Revision: https://reviews.llvm.org/D146001

ArgPromotion currently produces phantom / dead loads. A good example of this is store-into-inself.ll. First, ArgPromo finds the promotable argument %p in @l. Then it inserts a load of %p in the caller, and passes instead the loaded value / transforms the function body. PromoteMem2Reg is able to optimize out the entire function body, resulting in an unused argument. In a subsequent ArgPromotion pass, it removes the dead argument, resulting in a dead load in the caller. These dead loads may reduce effectiveness of other transformations (e.g. SimplifyCFG, MergedLoadStoreMotion).

This patch removes loads and geps that are made dead in the caller after removal of dead args.

Differential Revision: https://reviews.llvm.org/D146327

Differential revision: https://reviews.llvm.org/D146594

We should not call mdconst::extract, unless we know that the
metadata in question is ConstantAsMetadata.

For now we consider all other metadata as equal. The noalias test
shows that this is not correct, but at least it doesn't crash
anymore.

Summary:
The `exit` function in NVPTX has no intrinsic, but the assembly requires
a semicolon in the ptx, otherwise it will fail.

Memory fences are not handled by the NVPTX backend. We need to replace
them with a memory barrier intrinsic function. This doesn't include the
ordering, but should perform the necessary functionality, albeit slower.

Reviewed By: tianshilei1992

Differential Revision: https://reviews.llvm.org/D146725

Differential Revision: https://reviews.llvm.org/D144976

Switch from std::sort to std::stable_sort when sorting callsites to
avoid non-determinism when the comparisons are equal. This showed up in
internal testing of fe27495b.

In most cases the NOT will still be scalarized, but it allows us to perform the CMP(X,0) combines inside combineCMP()

I forgot to git add this test when committing the change.

For tests marked as "USE_SYSTEM_STDLIB", the expectation is that the
system's standard library should be used. However, the implementation of
this flag is such that we simply don't pass _any_ libcxxx-related flags
to Clang; in turn, Clang will use its defaults.

For a Clang/Libcxx pair compiled together, Clang defaults to:
1. The headers of the sibling libcxx.
2. The libraries of the system.

This mismatch is actually a bug in the driver; once fixed, however, (2)
would point to the sibling libcxx as well, which is _not_ what test
authors intended with the USE_SYSTEM_STDLIB flag.

As such, this patch explicitly sets a path to the system's libraries.
This change is done only in Apple platforms so that we can test this
works in this case first.

Differential Revision: https://reviews.llvm.org/D146714

This patch adds the OffloadEntriesInfoManager to the OpenMPIRBuilder, and
allows the OffloadEntriesInfoManager to access the Configuration in the
OpenMPIRBuilder.  With the shared Config there is no risk for inconsistencies,
and there is no longer the need for clang to have a separate
OffloadEntriesInfoManager.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D146549

We can simply push them down the existing call slowpath with some minor changes to how we compute the size argument.

[HWASAN] Disable unexpected_format_specifier_test because HWASAN doesn't provide a printf interceptor

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D146647

I noticed, when examining the generated Asm Matcher table, that some of
these custom immediate operands are missing, and so we are not parsing
some hint aliases into the correct MCInst.

Where this becomes apparent is when you parse e.g. `hint #7` into an
MCInst - without these cases, it becomes the MCInst `(HINT 17)`, which
will always be printed as `hint #17`. With these cases, it becomes the
MCInst `XPACLRI`, which will be printed as `xpaclri` with pauth, or
`hint #17` without, matching how `xpaclri` is parsed.

We only handle some specific hint aliases in this manner, usually where
these hints have specific effects that need to be modelled for accurate
code-generation. Otherwise, we just use the normal `InstAlias` system
to have the aliases parsed into a `(HINT N)` MCInst.

Differential Revision: https://reviews.llvm.org/D146630

Fix a regresion introduced by D124351.
Attributes of lambda call operator were evaluated in the
context of the closure object type rather than its operator,
causing an assertion failure.

This was because we temporarily switch to the class lambda to
produce the mangling of the lambda, but we stayed in that
context too long.

Reviewed By: eandrews, aaron.ballman

Differential Revision: https://reviews.llvm.org/D146535

This adds support for scalable vector types - at least far enough to get basic load and store cases working. It turns out that load/store without origin tracking already worked; I apparently got that working with one of the pre patches to use TypeSize utilities and didn't notice. The code changes here are required to enable origin tracking.

For origin tracking, a 4 byte value - the origin - is broadcast into a shadow region whose size exactly matches the type being accessed. This origin is only written if the shadow value is non-zero. The details of how shadow is computed from the original value being stored aren't relevant for this patch.

The code changes involve two related primitives.

First, we need to be able to perform that broadcast into a scalable sized memory region. This requires the use of a loop, and appropriate bound. The fixed size case optimizes with larger stores and alignment; I did not bother with that for the scalable case for now. We can optimize this codepath later if desired.

Second, we need a way to test if the shadow is zero. The mechanism for this in the code is to convert the shadow value into a scalar, and then zero check that. There's an assumption that this scalar is zero exactly when all elements of the shadow value are zero. As a result, we use an OR reduction on the scalable vector. This is analogous to how e.g. an array is handled. I landed a bunch of cleanup changes to remove other direct uses of the scalar conversion to convince myself there were no other undocumented invariants.

Differential Revision: https://reviews.llvm.org/D146157

Differential Revision: https://reviews.llvm.org/D144445
Reviewed By: nikic

fix: wrong arrow

Differential Revision: https://reviews.llvm.org/D145355

tweak: test

This patch fixes an issue whereby a constexpr class member which is
mapped to the device is being optimized out thus leading to a runtime
error.

Patch: https://reviews.llvm.org/D146552

Fix regression introduced by https://reviews.llvm.org/D123446

Reviewed By: tianshilei1992

Differential Revision: https://reviews.llvm.org/D146689

SVE and MTE both require a CPU with that feature before
you can use the other options, but we only added the "max"
cpu when SVE was enabled too.