This is similar to D112762, but the VL for the mask instruction
comes from a Phi node selecting between two vsetvlis.

The only binary-format-related field in the BBAddrMap structure is the function address (`Addr`), which will use uint64_t in 64B format and uint32_t in 32B format. This patch changes it to use uint64_t in both formats.
This allows non-templated use of the struct, at the expense of a marginal additional size overhead for the 32-bit format. The size of the BB address map section does not change.

Differential Revision: https://reviews.llvm.org/D112679

Rename vpopc/vmandnot/vmornot to vcpop/vmandn/vmorn assembler mnemonic.

Reviewed By: frasercrmck, jrtc27, craig.topper

Differential Revision: https://reviews.llvm.org/D111062

Reviewed By: bixia

Differential Revision: https://reviews.llvm.org/D113152

rdar://85020754

We generally prefer build bots which build each change as they come in as long as they can keep up. This generates much higher quality feedback to a developer. However, the code structure is such that we currently default new builders to batching.

This change updates the getting started instruction for a builder to explicitly advocate for not collapsing requests.

Differential Revision: https://reviews.llvm.org/D112776

Extend fusion on tensors to fuse producers greedily.

Reviewed By: nicolasvasilache, hanchung

Differential Revision: https://reviews.llvm.org/D110262

A default calculated size for SmallVector was added in
https://reviews.llvm.org/D92522 after discussion in
https://groups.google.com/g/llvm-dev/c/Z-VwNCTRGSg, but to_vector still
requires an explicit size. This patch adds the default size to to_vector
as well, so that this case doesn't unnecessarily force users to pick an
arbitrary size.

Reviewed By: silvas, dblaikie

Differential Revision: https://reviews.llvm.org/D112968

These relocations have been removed from the psABI doc in
https://github.com/riscv-non-isa/riscv-elf-psabi-doc/pull/199. This
patch simply updates llvm/include/llvm/BinaryFormat/ELFRelocs/RISCV.def
to match.

Differential Revision: https://reviews.llvm.org/D112999

On Armv6-M the branch may not able to reach the _Unwind_Resume function because it's relocation(R_ARM_THM_JUMP11) is in -2048, 2047 range only.

Reviewed By: chill, stuij, lenary

Differential Revision: https://reviews.llvm.org/D113181

In preparation for D112453.

Fix numbering of tests for [over.built]: 15->16, 16->17.

Summary:
Remove unused function from print-changed=dot-cfg code to silence a
gcc compiler warning.

Author: Jamie Schmeiser <schmeise@ca.ibm.com>
Reviewed By: uabelho(Mikael Holmen)
Differential Revision: https://reviews.llvm.org/D113188

This interface should not have existed in the first place, let alone
be a public member.

It allows calling `ElementCount::get(..)->getValue()`, which is ambiguous.
The interfaces to be used are either getFixedValue() or getKnownMinValue().

I noticed that, while go-to-def works on cases like:

namespace ns {
  template<typename T> struct Foo {};
}
using ::ns::Fo^o;

it only works because of the FileIndex. We can get definition location
directly from AST too.

Differential Revision: https://reviews.llvm.org/D113029

In a "Worksharing-loop construct", one can add a nowait clause at the end
of the loop (i.e. at the loop tail). This clause wasn't associated with
the corresponding loop when originally worked on in flang-omp-report.
Note that this refers to parsing and parse-tree generation. To work
around it, it was needed to track such clauses and the loops.
This should no longer be required (and in fact no longer works) and so
was removed. This results in 'curLoopLogRecord' and 'loopLogRecordStack' and
all references to them being removed. This also allows for
'constructClauses' to be swapped from std::deque to llvm::SmallVector.
Lastly a new test has been added testing the "nowait" clauses in a
variety of ways.

Reviewed By: awarzynski, kiranchandramohan

Differential Revision: https://reviews.llvm.org/D112217

Those nullptr checks are after we already accessed the pointer.

Reviewed By: labath

Differential Revision: https://reviews.llvm.org/D113175

https://reviews.llvm.org/D108392



Signed-off-by: Kevin Petit <kevin.petit@arm.com>

Function specialisation was running at all optimisation levels (if enabled on
the command line, it is not on by default). That was an oversight and not
something we want to do. Function specialisation duplicates functions when it
triggers, so the backend is processing more functions/instructions resulting in
compile-time increases, which seems more appropriate with -O3 and inline with
GCC. Please note that since function specialisation is not enabled by default,
this didn't require updating any pass manager tests.

Differential Revision: https://reviews.llvm.org/D112129

Reviewed By: labath

Differential Revision: https://reviews.llvm.org/D113176

This allows shipping individual functions without also having to provide
memset or bzero at the expense of bigger functions.

Similar to D113097.

Differential Revision: https://reviews.llvm.org/D113108

Similar to D113097 although not strictly necessary for now. It helps
keeping the same structure for all memory functions.

Differential Revision: https://reviews.llvm.org/D113103

This allows shipping individual functions without also having to provide
`memcpy` at the expense of bigger functions.
Next is to use this `inlined_memcpy` in:
 - loader/linux/x86_64/start.cpp
 - src/string/memmove.cpp
 - src/string/mempcpy.cpp
 - src/string/strcpy.cpp
 - src/string/strdup.cpp
 - src/string/strndup.cpp

Differential Revision: https://reviews.llvm.org/D113097

This patch removes unnecessary comments in the flang-omp-report
plugin tests which can be implied from the file name and run command.

Reviewed By: awarzynski

Differential Revision: https://reviews.llvm.org/D112817

Ensure that the destination slot exists in this case. This addresses PR51484.

rG1e5f8143 added the test case, I've added PR52397 to the comment to help keep track of the source of the bug

Check both operands for zero elements to remove unnecessary demanded elts.

Try to help reduce some minor regressions noticed in D110995

CPSR on Arm64 is 4 bytes in size but windows on Arm implementation is trying to read/write 8 bytes against a byte register causing LLDB unit tests failures.

Ref: https://docs.microsoft.com/en-us/windows/win32/api/winnt/ns-winnt-arm64_nt_context

Reviewed By: mstorsjo

Differential Revision: https://reviews.llvm.org/D112471

LIT skips various system environment variables while building test
config. It turns out that we require PLATFORM environment variable for
detection of x86 vs Arm windows platform.

This patch adds system environment variable PLATFORM into LIT test
config for detection of win32 Arm platform.

Reviewed By: mstorsjo

Differential Revision: https://reviews.llvm.org/D113165

As seen in https://bugs.llvm.org/show_bug.cgi?id=52213 llvm-objdump
asserts if either the --debug-vars or the --dwarf options are provided
with invalid values. As suggested, this fix adds use of a default value
to these options and errors when given bad input.

Differential Revision: https://reviews.llvm.org/D112183

Coroutines have weird semantics that don't quite match normal LLVM functions,
so trying to infer even simple attributes based on thier contents can go wrong.

Convert fir.insert_on_range operation to corresponding
llvm.insertvalue operations.

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: mehdi_amini

Differential Revision: https://reviews.llvm.org/D112896

The added test has poison lanes due to the vector shuffle. This can
cause an infinite loop of combines in instcombine where it folds
xor(ashr, -1) -> select (icmp slt 0), -1, 0 -> sext (icmp slt 0) -> xor(ashr, -1).
We usually prevent this by checking that the xor constant is not -1,
but with vectors some of the lanes may be -1, some may be poison. So
this changes the way we detect that from "!C1->isAllOnesValue()" to
"!match(C1, m_AllOnes())", which is more able to detect that some of the
lanes are poison.

Fixes PR52397

These are not standard methods, neither libstdc++ nor MSVC STL provide
them.

In practice, one of them was untested and the other one was only used in
one single test.

Differential Revision: https://reviews.llvm.org/D113027