Improve O(N^2) to O(N) in some cases, reduce number of allocations by
reserving memory.
Also, improve analysis of loads reduction values to avoid analysis
of not vectorizable cases.

Differential Revision: https://reviews.llvm.org/D136619

The following sequence should be folded into in0 * in1
      In0Lo = in0 & 0xffffffff; In0Hi = in0 >> 32;
      In1Lo = in1 & 0xffffffff; In1Hi = in1 >> 32;
      m01 = In1Hi * In0Lo; m10 = In1Lo * In0Hi; m00 = In1Lo * In0Lo;
      addc = m01 + m10;
      ResLo = m00 + (addc >> 32);

Reviewed By: spatel, RKSimon
Differential Revision: https://reviews.llvm.org/D136015

This reverts commit 65aaecca.

There was an ordering problem in the calculation of the partial
remainder.

Original commit message:

If the divisor is even, we can first shift the dividend and divisor
right by the number of trailing zeros. Now the divisor is odd and we
can do the original algorithm to calculate a remainder. Then we shift
that remainder left by the number of trailing zeros and add the bits
that were shifted out of the dividend.

Differential Revision: https://reviews.llvm.org/D135541

Address comments in D136372

Reviewed By: aartbik, springerm

Differential Revision: https://reviews.llvm.org/D136510

The test TestObjCDirectMethods loads the Objective C runtime, which
doesn't work well with custom a libcxx, resulting in two copies of the
standard library being loaded at runtime.

Like what was done for `TestObjCExceptions`, this commit forces the
usage of the system's library instead. The minimum required Clang
version is set to the oldest Clang that can compile the libraries
available in the lldb-matrix bots.

Differential Revision: https://reviews.llvm.org/D136600

This requirement dates back to ten years ago and the test seems to work
nowadays with either libc++ or libstdc++.

Differential Revision: https://reviews.llvm.org/D136608

Differential Revision: https://reviews.llvm.org/D136607

Implement P2513

This change allows initializing an array of unsigned char,
or char from u8 string literals.
This was done both to support legacy code and for compatibility
with C where char8_t will be typedef to unsigned char.

This is backported to C++20 as per WG21 guidance.

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D136449

This teaches the SCCP Solver how to constant fold more intrinsics. Constant
folding appears to be just as good as D115737 but much, much lower in code
change impact as suggested by nikic.

The constrained floating-point intrinsics all take at least one metadata
argument and were the motivation for the change.

Differential Revision: https://reviews.llvm.org/D136466

Differential Revision: https://reviews.llvm.org/D132385

Introduces walkUsed, a very simple version of the public API to enable
incremental development on rest of the pieces.

Differential Revision: https://reviews.llvm.org/D136293

(sign|resign) + (auth|resign) can be folded by omitting the middle
sign+auth component if the key and discriminator match.

Differential Revision: https://reviews.llvm.org/D132383

When lowering vector shuffles into the xxsplti32dx instruction on Power10, we
canonicalize the right operand to be a BUILD_VECTOR and as a result, get the
commuted vector shuffle node.

However, a vector shuffle will not always be returned as the result for a
commuted vector shuffle. In such a scenario, this patch updates the original
cast of a shuffle into a dyn_cast<> and checks if the shuffle is a valid vector
shuffle node prior to obtaining the commuted shuffle mask.

This patch also adds a new test case that demonstrates this scenario (primarily
seen on 32-bit), and was originally a crash prior to this fix.

Differential Revision: https://reviews.llvm.org/D135024

This reverts commit f6a7b478.

I received a report that this fails on 32-bit X86.

Differential Revision: https://reviews.llvm.org/D136370

Build gtest targets when building standalone only if LLDB_INCLUDE_TESTS
is true.  Prior to this change, they were built whenever
LLVM_MAIN_SRC_DIR was available, independently whether tests themselves
would be run.

Differential Revision: https://reviews.llvm.org/D136551

This appears to be a copy+paste typo in the znver1/2 AMD SoG tables, treating the byte shift instructions like bit shifts

Older AMD SoG referred to PSLLDQ/PSRLDQ as shuffles, and Agner/instlatx64 both report they are integer shuffles

This patch adds the kernel to lower evaluate::Expr to HLFIR to a
hlfir::FortranEntity (a single mlir::Value that can be interpreted as
a Fortran variable or the value of a Fortram expression).

It implements lowering of simple name designators ("x") and starts
adding a translation layer in AbstractConverter::genExprBox and
AbstractConverter::genExprAddr so that the new expression lowering
can be used without any changes for now in the current statement and
construct lowering.

Differential Revision: https://reviews.llvm.org/D136453

As that bug reports, the problem here is that the lambda's
'context-decl' was not set to the concept, and the lambda picked up
template arguments from the concept.  SO, we failed to get the correct
template arguments in SemaTemplateInstantiate.

However, a Concept Specialization is NOT a decl, its an expression, so
we weren't able to put the concept in the decl tree like we needed.
This patch introduces a ConceptSpecializationDecl, which is the smallest
type possible to use for this purpose, containing only the template
arguments.

The net memory impliciation of this is turning a
trailing-objects into a pointer to a type with trailing-objects,  so it
should be minor.

As future work, we may consider giving this type more responsibility, or
figuring out how to better merge duplicates, but as this is just a
template-argument collection at the moment, there isn't much value to
it.

Differential Revision: https://reviews.llvm.org/D136451

Try to get better coverage for the pattern-matching
possibilities in D136015.

This gets the tests closer to the form that we are
trying to test in D136015. Note that the IR has
changed, but the check lines have not changed.

This also shows that the desired commuted pattern
coverage is not as expected.

When checking for non null arguments the wrong SourceLocation was given,
this fix to pass the proper argument's location.

Fixes #58273

Differential Revision: https://reviews.llvm.org/D136355

There was a bug in scf.for loop bufferization that could lead to a missing buffer copy (alloc was there, but not the copy).

Differential Revision: https://reviews.llvm.org/D135053

This adds Clang command line support for the NVIDIA Grace CPU [1], which
we would like to target with -mcpu=grace.

Grace is based on the Arm Neoverse V2 CPU [1], which is why, in the
driver, we pass neoverse-v2 to the compiler when grace is requested.
Thus, for now, it's an alias to Neoverse V2.

This adds support for grace in user-facing tool Clang, but not yet the
developer tools like opt and llc, so we probably want to follow up on
this.

[1] https://www.nvidia.com/en-us/data-center/grace-cpu/

Differential Revision: https://reviews.llvm.org/D136425

Recognize when sub-vectors have been split to elements which are used to
build large vector.
This happens when instructions have different vector sizes available.
For example a few arithmetic instruction are required to process all
elements of larger vector that can be stored using one instruction.

Differential Revision: https://reviews.llvm.org/D109242

Recognize when source could have been unmerged to pieces with DstTy
without having to split source to smaller elements
and then merge small elements into DstTy pieces.
This happens when vector was meant to be split to sub-vectors but there
was leftover. At this point artifact combiner have already dealt with
leftover and we can continue to use sub-vectors.

Differential Revision: https://reviews.llvm.org/D109241

Recognize copy that is represented as split of a source register to
elements that were reassembled to another register with the same type.

Differential Revision: https://reviews.llvm.org/D109240

Differential Revision: https://reviews.llvm.org/D117655

The lack of this attribute (particularly on the load intrinsics)
prevented InstCombine from optimizing away allocas and memcpys
for arrays that could be read directly from rodata.

This now also includes a new test to check the masked load/store
intrinsics have the expected attributes (specifically nocapture).

Differential Revision: https://reviews.llvm.org/D135656

This was broken by https://reviews.llvm.org/D136351

`cpp::byte` is better than `char` which -depending on platform- can be `signed char` or `unsigned char`.  This has introduced subtle arithmetic errors.

This switches the v8.5-a FRINT intrinsics over to be target-gated,
behind preprocessor defines. This one is pretty simple, being AArch64
only.

Differential Revision: https://reviews.llvm.org/D135646

Move include_directories() declaration before gtest targets are created
in standalone build.  This fixes build failure due to gtest targets
being unable to find LLVM headers, e.g.:

    /var/tmp/portage/dev-util/lldb-16.0.0_pre20221023/work/llvm/utils/unittest/googletest/include/gtest/internal/custom/raw-ostream.h:43:10: fatal error: llvm/ADT/Optional.h: No such file or directory

Differential Revision: https://reviews.llvm.org/D136552