Also moved the range from the function's call sites to the functions
return value as that is possible now.

This reverts commit b77416ea. While that seemed like an
appropriate rename at the time, further review on pull request #90078
makes it unnecessary. This revert changes the name back.

Currently, the OptimizePointerDifference fold does not trigger when
working on the sub of two geps where one of the geps has multiple uses,
to avoid duplicating the offset arithmetic too much.

However, there are cases where performing it would still be
clearly profitable, e.g. test_sub_ptradd_multiuse.

This patch drops the one-use restriction using the same strategy we use
in GEP comparison folds: If there are multiple uses, we rewrite the GEP
to use the expanded offset arithmetic instead (effectively
canonicalizing it into ptradd representation).

Fixes https://github.com/llvm/llvm-project/issues/88231.

Compare the predecessor's exit after we've already checked that the PHI
comes from a vsetvli. That way it's more obvious that Require has the same
VL as PBBExit.

Canonicalization defaulted to replacement when the input dims were from
unknown source. This is obviously incorrect. Tweaked and included test
to prevent future issue.

Allow using common::visit with Clang.

Test plan: ninja check-all

Instead of manually checking hasSameVTYPE and hasSameAVL.

Revert "[Clang][Sema] Diagnose class member access expressions naming non-existent members of the current instantiation prior to instantiation in the absence of dependent base classes (#84050)"

This reverts commit a8fd0d02.

This reverts commit 6dd2617c.

`useInBoundsInsteadOfMasking` was doing the opposite i.e., when set to
true; was updating the mask instead of updating the inBounds.

Keep the DemandedFields logic grouped together in the struct

Jonas upstreamed recognition of DW_TAG_LLVM_ptrauth_type
https://reviews.llvm.org/D130215 but it isn't recognized as a type
qualifier tag in DWARFASTParserClang::ParseTypeFromDWARF.

I previously added this API via https://reviews.llvm.org/D142792 in
2023, along with changes to the ValueObject class to treat pointer types
as addresses, and to annotate those ValueObjects with the original
uint64_t byte sequence AND the name of the symbol once stripped, if that
points to a symbol.

I did this unconditionally for all pointer type ValueObjects, and it
caused several regressions in the Objective-C data formatters which have
a ValueObject of an object, it has the address of its class -- but with
ObjC, sometimes it is a "tagged pointer" which is metadata, not an
actual pointer. (e.g. a small NSInteger value is stored entirely in the
tagged pointer, instead of a separate object) Treating these
not-addresses as addresses -- clearing the non-addressable-bits -- is
invalid.

The original version of this patch we're using downstream only does this
bits clearing for pointer types that are specifically decorated with the
pointerauth typequal, but not all of those clang changes are upstreamed
to github main yet, so I tried this simpler approach and hit the tagged
pointer issue and bailed on the whole patch.

This patch, however, is simply adding SBValue::GetValueAsAddress so
script writers who know that an SBValue has an address in memory, can
strip off any metadata. It's an important API to have for script writers
when AArch64 ptrauth is in use, so I'm going to put this part of the
patch back on github main now until we can get the rest of that original
patch upstreamed.

This patch covers a really basic reassociation optimizations for VADD_VV and VMUL_VV.

With the tag merging in place, we can safely change the default for
+seq-cst-trailing-fence to the default, according to the recommendation
in

https://github.com/riscv-non-isa/riscv-elf-psabi-doc/blob/master/riscv-atomic.adoc

This tag changes the default for the feature flag, and moves to more
consistent naming with respect to existing features.

This patch adds support for the `atomic_abi` attribute, specifid in

https://github.com/riscv-non-isa/riscv-elf-psabi-doc/blob/master/riscv-elf.adoc#tag_riscv_atomic_abi-14-uleb128version.

The atomics_abi tag merging is conducted as follows:

- UNKNOWN is safe to merge with all other values.
- A6C is compatible with A6S, and results in the A6C ABI.
- A6C is incompatible with A7, and results in an error.
- A6S and A7 are compatible, and merging results in the A7 ABI.

Note: the A7 is not yet supported in either LLVM or in any current
hardware,
and is therefore ommited from attribute generation in
RISCVTargetStreamer.

[SLP]Fix PR89988: do extra analysis of the icmp args to correctly handle signed/unsigned comparison.

If operands of icmp has different signedness, need to consider extending
unsigned operands to correctly handle comparison with the signed
operands.

This is the pre-commit test for PR #88307.

Copy of llvm/test/CodeGen/X86/mmx-intrinsics.ll

This patch replaces count with contains, following the spirit of
clang-tidy's readability-container-contains.

Previously, the memprof and ctx_profile lit.site.cfg would assert
if the enable_per_target_runtime_dir was set and the
config.target_arch != config.host_arch. However, config.host_arch would
never be set, meaning this would always fail in a when using
LLVM_ENABLE_PER_TARGET_RUNTIME_DIR.

This patch follows the example in the ASAN lit.site.cfg.py that updates
the compiler_rt_libdir appropriately.

Completely refactor the cache file to simplify it and remove unnecessary
variables. The main functional change here is that the non-PGO builds
now use two stages, so `ninja -C build stage2-package` can be used with
both PGO and non-PGO builds.

@jimingham I am wondering if you are ok removing this test? It caused
failures in some of the build bots because the user time was less than a
microsecond. Alternatively we can increase the number of loops or maybe
I need some other approach? I had commented it out just to not impact
others

Add the callsite intrinsic. 

Structurally, it is very similar to the counter intrinsics, hence the inheritance relationship. We can probably rename `InstrProfCntrInstBase` to `InstrProfIndexedBase` later - because the "counting" aspect is really left to derived types of `InstrProfCntrInstBase`, and it only concerns itself with the index aspect (which is what we care about for `callsite`, too)

(Tracking Issue: #89287, RFC referenced there)

Related to clean-up opportunities discussed at #90013.

After these cleanups, the `RelaxAll` parameter from
`createMCObjectStreamer` can be removed as well. As
`createMCObjectStreamer` is a more user-facing API and used by two files
in mlir/, we postpone the cleanup to the future.

Use [[fallthrough]] directly (similarly to
https://reviews.llvm.org/D131346).

This recommits eb05a2e8.

By expressing the conditions covered by MAP_ALLOWNOMEM in a more
direct way.

To correctly lower multi_reduction of 1-dim scalable vector, e.g., <[4]xf32>

Related to clean-up opportunities discussed at #90013.

Added GISelShouldIgnore property to class Pattern in TargetSelectionDAG.td; it's similar to FastISelShouldIgnore. This bit can be put on a record to avoid its pattern import within GlobalISelEmitter. This allows one to avoid the record's GISel .td implementation, .inc generation, and any skipped pattern warnings from -warn-on-skipped-patterns.

Extend LoopAccessAnalysis to support different strides and as a
consequence non-constant distances between dependences using SCEV to
reason about the direction of the dependence.

In multiple places, logic to rule out dependences using the stride has
been updated to only be used if StrideA == StrideB, i.e. there's a
common stride.

We now also may bail out at multiple places where we may have to set
FoundNonConstantDistanceDependence. This is done when we need to bail
out and the distance is not constant to preserve original behavior.

Fixes https://github.com/llvm/llvm-project/issues/87336

PR: https://github.com/llvm/llvm-project/pull/88039

The preprocessor define `__HLSL_ENABLE_16_BIT` should be set to 1 if
native 16-bit types are enabled and not set if they are not.

Previously we were setting the value to match the HLSL active language
version, and we had no test coverage verifing the value was set and not
set as expected.

Fixes #89787

Adds a test for `inferConvolutionDimsImpl` to exercise the logic for
depthwise convs.