Inserting scalar result of 'uaddlv' neon intrinsic function to a destination vector currently makes use of the integer unit. Subsequent patches will eliminate the redundant use of the integer registers in a more generic way that will include this special case. This is an initial set of tests for this functionality.

Differential Revision: https://reviews.llvm.org/D143038

We use `use64bitIndex` in the option to decide the target device
address bitwidth. This makes it consistent with index type
conversion too.

Reviewed By: kuhar

Differential Revision: https://reviews.llvm.org/D144827

Vulkan requires GPU processor ID/count builtin variables to be
32-bit scalar or vector for all the cases. Similarly there
are special requirements for OpenCL. We need to make sure those
rules are respected when converting using 64bit for index.

Reviewed By: kuhar

Differential Revision: https://reviews.llvm.org/D144819

The asm printer grew the ability to automatically fall back to the
generic format for invalid ops, so this logic doesn't need to be in the
bindings anymore. The printer already handles supressing diagnostics
that get emitted while checking if the op is valid.

Reviewed By: mehdi_amini, stellaraccident

Differential Revision: https://reviews.llvm.org/D144805

Signed-off-by: Jun Zhang <jun@junz.org>

If a template function contained a pragma that made it strictfp, code
generation for such function crashed, because the instantiation did not
have strictfp attribute. As a solution this attribute is copied from the
template to instantiation.

Differential Revision: https://reviews.llvm.org/D143919

/data/llvm-project/llvm/lib/Target/X86/MCTargetDesc/X86MCCodeEmitter.cpp:1200:3: error: unannotated fall-through between switch labels [-Werror,-Wimplicit-fallthrough]
  case X86II::RawFrm:
  ^

1. Add a variable `HasRegOp` to record if the instruction has a register operand
2. Enumerate all the formats with a register operand in the switch
2. Add a default (unreachable) label in the switch (suggested by @reames)

Reviewed By: pengfei

Differential Revision: https://reviews.llvm.org/D144776

We use this combine in the AArch64 postlegalizer combiner, which causes this
function to query the legalizer rules for the action for an invalid opcode/type
combination (G_AND and p0). Moving the legalizer query until after the validity
check in matchHoistLogicOpWithSameOpcodeHands() fixes this.

This patch implements the exception handling rules found in
N4849 14.4 and as a result fixes many false positives in
the bugprone-exception-escape checker.

Differential Revision: https://reviews.llvm.org/D135495

This commit just adds options to control index type bitwidth in
GPUToSPIRV conversion, and updates tests to prepare for 64bit
index conversion.

Reviewed By: kuhar

Differential Revision: https://reviews.llvm.org/D144826

The `tyname_cell` definitions at the top are now all the same, so
replace them with a single `alloca_cell` type.

Preparation for further additions.

This patch adds support of DWARFv5 .debug_rnglists table.
As DWARFLinker resolves relocations, it is able to always
use DW_FORM_addr instead of DW_FORM_addrx. DW_FORM_addrx
helps to minimize number of relocations, it is also used for
split DWARF. Both of these cases are not relevant for the
DWARFLinker. Thus, this patch converts all DW_FORM_addrx
forms into the DW_FORM_addr. And, as the result, it converts
range lists of DW_FORM_rnglistx form into the DW_FORM_sec_offset.
For the --update case all DW_FORM_addrx, DW_FORM_rnglistx
are preserved as is.

Reviewed By: aprantl

Differential Revision: https://reviews.llvm.org/D143903

Added: OR, SMAX, SMIN, UMAX, UMIN, ABS, SHL, SRL, SRA, MUL

Intentionally not generically using TLI.isBinOp as that causes
regressions as there are many binops that cannot combine with masked
instructions.

Reviewed By: RKSimon

Differential Revision: https://reviews.llvm.org/D143860

Differential Revision: https://reviews.llvm.org/D144144

`{v}blend{d|ps|pd}` is preferable to `{v}blendw` so widen so that we
can match it.

Reviewed By: RKSimon

Differential Revision: https://reviews.llvm.org/D143789

Reviewed By: RKSimon

Differential Revision: https://reviews.llvm.org/D143788

Invariant.group's are not sufficiently handled by LICM. Specifically,
if a given invariant.group loaded pointer is not overwritten between
the start of a loop, and its use in the load, it can be hoisted.
The invariant.group (on an already invariant pointer operand) ensures
the result is the same. If it is not overwritten between the start
of the loop and the load, it is therefore legal to hoist.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D144053

Fix https://github.com/llvm/llvm-project/issues/59740.
NOTE: retry as we can't reproduce the break locally when first commit.

Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D141188

bitselect(52,-6553,m) -> xor(and(xor(52,-6553),m),52) folds much better than or(and(52,not(m)),and(-6553,m))

As noted on Issue #45817 we didn't have scalar coverage for this

Implemented support for bit-field members as a loop variable
or upper limit. Supporting also non bit-field integer members.

Fixes issues: https://github.com/llvm/llvm-project/issues/58614

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D142587

The flag name isn't the best: avoiding negative is in general more readable
and conveys the intent better.

Extend tests for D144449 to ensure we aren't regressing code on targets with a slow INC/DEC instruction (or i686 without CMOV)

A somewhat common code-pattern is to default a destructor in the source file and not in the header.
For example, this is the way to use smart pointers with forward-declared classes:

```c++

struct Impl;
struct A {
  ~A(); // Can't be defaulted in the header.

private:
  std::unique_ptr<Impl> impl;
};
```

To be able to use this check with this pattern, I modified the behavior with `AllowSoleDefaultDtor`
to not trigger on destructors if they aren't defined yet.
Since a declared destructor should still be defined somewhere in the program, this
won't miss bad classes, just diagnose on less translation units.

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D143851

The current clang-tidy settings work in the CI but not on all systems
outside the CI.
- The range 16...17 doesn't work when only clang-17 is installed.
- Running CMake a second time will fail.

This addresses these issues.

Reviewed By: philnik, #libc

Differential Revision: https://reviews.llvm.org/D144785

We don't want to produce inlay hints for arguments for which
user has left param name comments. But we're not decomposing
location of the parameter correctly at the moment because the
location we've passed into `SM.getDecomposedLoc` is not always
FileID.

Fixes clangd/clangd#1495

Reviewed By: nridge

Differential Revision: https://reviews.llvm.org/D144074

Calling the runtime on disassociated pointer or unallocated
allocatable will trigger a segfault.

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D144752