[AMDGPU] Skip debug instruction uses while optimizing live range of a reg in SIOptimizeVGPRLiveRange

This will prevent the `assert(!O.readsReg())` from firing in
SIOptimizeVGPRLiveRange::optimizeLiveRange

Fix for #64163

Reviewed By: arsenm, #amdgpu

Differential Revision: https://reviews.llvm.org/D156893

[RISCV] Remove -menable-experimental-extensions option for zc extension in riscv-target-features.c. NFC.

zca, zcb, zcd, zcf, zcmp and zcmt aren't experimental extensions.

This commit introduces a debug location verifier for the LLVM dialect's
call operation. LLVM does not allow calls to have no debug location when
they reference an inlinable function with debug information. This
apparenlty breaks assumptions of LLVM's inliner.

So far, there was a hack in the LLVM export that avoided this case to
be triggered, but that hack causes issues when debug intrinsics are
involved. Link to the revision that inroduced the export hack:
https://reviews.llvm.org/D88135

LLVM's verifier as a reference: https://github.com/llvm/llvm-project/blob/2df05cd01c17f3ef720e554dc7cde43df27e5224/llvm/lib/IR/Verifier.cpp#L3546

Reviewed By: gysit, zero9178

Differential Revision: https://reviews.llvm.org/D157096

abs(x-y) --> x-y where x >= y, done on D122013
abs(x-y) --> y-x where x <= y

proofs: https://alive2.llvm.org/ce/z/KkeEsd

Reviewed By: goldstein.w.n, nikic
Differential Revision: https://reviews.llvm.org/D156499

Reviewed By: goldstein.w.n
Differential Revision: https://reviews.llvm.org/D156521

Some downstream clang is configured to add -rpath to HIP runtime for ld.
Also clang may use code object version by default.

Make this test result not affected by those.

Reviewed by: Joseph Huber

Differential Revision: https://reviews.llvm.org/D157243

Replace ConditionalAssume set by treating conditional assumes like other
predicated instructions (i.e. create a VPReplicateRecipe with a mask)
and later remove any assume recipes with masks during VPlan cleanup.

This reduces coupling of VPlan construction and Legal by removing a
shared set between the 2 and results in a cleaner code structure
overall.

Reviewed By: Ayal

Differential Revision: https://reviews.llvm.org/D157034

Fixes #63519.

Differential Revision: https://reviews.llvm.org/D157179

This now reuses the existing lowering for G_FMIN/MAX for G_FABS too, which can
handle more type successfully. We can hopefully reuse the same pattern action
definition for other fp operations too.

This replicates the G_FMINNUM and G_FMAXNUM lowering to G_FMINIMUM and
G_FMAXIMUM, reusing the same action definition for lowering.

Now covers gisel as well as selection dag, and more types are tested. The
existing tests for combines to fabs are moved to fabs-combine.ll.

This patch modifies the plugins so that the initialization of KernelTy objects
is done in the init method. Part of the initialization was done in the
constructKernelEntry method. Now this method is called constructKernel
and only allocates and constructs a KernelTy object.

This patch prepares the kernel class for the new implementation of device
reductions.

Differential Revision: https://reviews.llvm.org/D156917

Keep release notes in same format, remove the.

Enforce a stricter match with the swap function signature, eliminating false-positives.

Fixes: #64303

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D157185

Issue were cased by checking if canonical declaration of record is a
definition. Unfortunately when forward declarations were used, it were
not a definition.

Fixes: #64463

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D157190

Looks like somewere in clang-16 this alias were removed,
even that is mention in clang-tidy 15 release notes.

Fixes: #64342

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D157181

Broke sanitizer buildbot:
https://lab.llvm.org/buildbot/#/builders/240/builds/12947
https://lab.llvm.org/buildbot/#/builders/19/builds/18369

Reverting while I figure out the cause.

This reverts commit bd0ed0ab.

This patch marks the declarations with initializations in condition expressions such as
if (int var = init) as unreferenced so that -Wunused can warn on them.

Fixes https://github.com/llvm/llvm-project/issues/61681

Reviewed By: cor3ntin
Differential Revision: https://reviews.llvm.org/D152495

Fix link to cppcoreguidelines-avoid-goto in doc.

This is an alternative to D156967 where I suggested the author
could use a VBR type.

This patch takes inspirations from Opc_EmitRegister2 that is used
for two byte registers.

I'm assuming 1 or 2 byte predicates should be enough so we don't
need the fully generality of VBR.

This avoids impacting the table size on targets that have more than
128 predicates already like X86.

Reviewed By: bogner

Differential Revision: https://reviews.llvm.org/D157196

Rather than have distinct implementations of memcpy.inline and memmove
creators, just use the existing CreateMemTransferInst, which plain
memcpy already uses. The AMD backend already generates memmove from
that function.

Reviewed By: gchatelet, jroelofs

Differential Revision: https://reviews.llvm.org/D156991

AutoUpgrade detection of memcpy and memmove is almost
identical. Commonize the code path.

Reviewed By: jroelofs

The modernize-loop-convert check will now match iterator based loops
that call the free functions 'begin'/'end', as well as matching the
free function 'size' on containers.

Test plan: Added unit test cases matching free function calls on
containers, and a single negative test case for length() which is not
supported.

Reviewed By: PiotrZSL

Differential Revision: https://reviews.llvm.org/D140760

The last use was removed by:

  commit 5a7c2f17
  Author: Eric Christopher <echristo@gmail.com>
  Date:   Thu Jun 22 22:58:12 2017 +0000

Fixes 63393.

Differential Revision: https://reviews.llvm.org/D156971

[clang-tidy] Exclude class/struct scope variables from cppcoreguidelines-avoid-non-const-global-variables

Ignore static variables declared within the scope of class/struct.
Those variables should be covered by I.3 rule.

Fixes: #47384

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D157180

Add info in check list that hicpp-avoid-goto is actually
a alias to cppcoreguidelines-avoid-goto.

Fixes: #64337

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D157182

Provide fixes for inline namespaces in the same format as clang-format.

Fixes: #56804

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D157178

[X86] Only fold broadcast with extract_vector_elt/scalar_to_vector if the scalar type matches the vector element type

Avoid handling implicit extension/truncation with scalar<->vector transfers

Fixes #64439

Extend the existing 128-bit -> 64-bit truncation handling by widening/narrowing the src/dst vectors and use the lower half operand/result for PACKSS/PACKUS instructions.

Fixes #29973

The original test has a unused load, which is removed. Also add a
variant with a store that cannot be removed, forcing the mask for the
block to always be generated.

This patch adds a lightweight instance of Attributor that only deduces
attributes.

This is just an initial version with the goal to have a version that
only focuses on attributes to replace the function-attrs pass.

The initial version has a few open issues pending until default
enablement, the main one probably being compile time. The main
additional functionality this will provide in general is propagating
attributes to call sites.

Open issues:

* compile time
  The current version increase O3 +2.67% and ThinLTO +6.18% when replacing FunctionAttr
   https://llvm-compile-time-tracker.com/compare.php?from=c4bb3e073548cf436d5fa0406e3ae75e94684dec&to=d992630a69c79a2587d736e6a88f448850413bd1&stat=instructions%3Au
   Both are with an additional change to preserve more analysis, like FunctionAttrs CGSCC run.

* some missed attribute inference

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D152081

This stops tablegen from generating unneeded entries in
RISCVGenDAGISel.inc.

Enable record-replay for OpenMP offload kernels.  On recording the initialization
is performed on device initialization by reading env variables. (This is similar to
the way rr used to operate). The primary change takes place in the replay phase
with the replay tool explicitly initializing the record-replay functionality.

Differential Revision: https://reviews.llvm.org/D156174

Fix

The last use was removed by:

  commit b14e30f1
  Author: Sameer Sahasrabuddhe <sameer.sahasrabuddhe@amd.com>
  Date:   Thu Jul 27 09:54:50 2023 +0530