Changes

[WebAssembly] Optimize out shift masks

WebAssembly's shift instructions implicitly masks the shift count, so optimize
out redundant explicit masks of the shift count. For vector shifts, this
currently only works if the mask is applied before splatting the shift count,
but this should be addressed in a future commit. Resolves PR49655.

Differential Revision: https://reviews.llvm.org/D105600

[WebAssembly][lld] Fix segfault on .bss sections in mapfile

When memory is declared in the Wasm module, we rely on the implicit zero
initialization behavior and do not explicitly output .bss sections. The means
that they do not have associated `outputSec` entries, which was causing
segfaults in the mapfile support. Fix the issue by guarding against null
`outputSec` and falling back to using a zero offset.

Differential Revision: https://reviews.llvm.org/D102951

[ORC] Improve computeLocalDeps / computeNamedSymbolDependencies performance.

The computeNamedSymbolDependencies and computeLocalDeps methods on
ObjectLinkingLayerJITLinkContext are responsible for computing, for each symbol
in the current MaterializationResponsibility, the set of non-locally-scoped
symbols that are depended on. To calculate this we have to consider the effect
of chains of dependence through locally scoped symbols in the LinkGraph. E.g.

        .text
        .globl  foo
foo:
        callq   bar                    ## foo depneds on external 'bar'
        movq    Ltmp1(%rip), %rcx      ## foo depends on locally scoped 'Ltmp1'
        addl    (%rcx), %eax
        retq

        .data
Ltmp1:
        .quad   x                      ## Ltmp1 depends on external 'x'

In this example symbol 'foo' depends directly on 'bar', and indirectly on 'x'
via 'Ltmp1', which is locally scoped.

Performance of the existing implementations appears to have been mediocre:
Based on flame graphs posted by @drmeister (in #jit on the LLVM discord server)
the computeLocalDeps function was taking up a substantial amount of time when
starting up Clasp (https://github.com/clasp-developers/clasp).

This commit attempts to address the performance problems in three ways:

1. Using jitlink::Blocks instead of jitlink::Symbols as the nodes of the
dependencies-introduced-by-locally-scoped-symbols graph.

Using either Blocks or Symbols as nodes provides the same information, but since
there may be more than one locally scoped symbol per block the block-based
version of the dependence graph should always be a subgraph of the Symbol-based
version, and so faster to operate on.

2. Improved worklist management.

The older version of computeLocalDeps used a fixed worklist containing all
nodes, and iterated over this list propagating dependencies until no further
changes were required. The worklist was not sorted into a useful order before
the loop started.

The new version uses a variable work-stack, visiting nodes in DFS order and
only adding nodes when there is meaningful work to do on them.

Compared to the old version the new version avoids revisiting nodes which
haven't changed, and I suspect it converges more quickly (due to the DFS
ordering).

3. Laziness and caching.

Mappings of...

jitlink::Symbol* -> Interned Name (as SymbolStringPtr)
jitlink::Block* -> Immediate dependencies (as SymbolNameSet)
jitlink::Block* -> Transitive dependencies (as SymbolNameSet)

are all built lazily and cached while running computeNamedSymbolDependencies.

According to @drmeister these changes reduced Clasp startup time in his test
setup (averaged over a handful of starts) from 4.8 to 2.8 seconds (with
ORC/JITLink linking ~11,000 object files in that time), which seems like
enough to justify switching to the new algorithm in the absence of any other
perf numbers.

[ORC] Introduce ExecutorAddress type, fix broken LLDB bot.

ExecutorAddressRange depended on JITTargetAddress, but JITTargetAddress is
defined in ExecutionEngine, which OrcShared should not depend on.

This seems like as good a time as any to introduce a new ExecutorAddress type
to eventually replace JITTargetAddress. For now it's just another uint64_t
alias, but it will soon be changed to a class type to provide greater type
safety.

[mlir][linalg] Tighter StructuredOp Verification.

Verify the number of results matches exactly the number of output tensors. Simplify the FillOp verification since part of it got redundant.

Differential Revision: https://reviews.llvm.org/D105427

[lld/mac] Fix warning about unused variable [NFC]

Change "dyn_cast" to "isa" to get rid of the unused
variable "bitcodeFile".

gcc warned with

lld/MachO/Driver.cpp:531:17: warning: unused variable 'bitcodeFile' [-Wunused-variable]
531 |       if (auto *bitcodeFile = dyn_cast<BitcodeFile>(file)) {
    |                 ^~~~~~~~~~~

[AMDGPU] Fix typo

[mlir] factor out ConvertToLLVMPattern

This class and classes that extend it are general utilities for any dialect
that is being converted into the LLVM dialect. They are in no way specific to
Standard-to-LLVM conversion and should not make their users depend on it.

Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D105542

[mlir][MemRef] Fix DimOp folding of OffsetSizeAndStrideInterface.

This addresses the issue reported in

https://llvm.discourse.group/t/rank-reducing-memref-subview-offsetsizeandstrideopinterface-interface-issues/3805

Differential Revision: https://reviews.llvm.org/D105558

[mlir][linalg] Remove GenericOpBase.

Remove the GenericOpBase class formerly used to factor out common logic shared be GenericOp and IndexedGenericOp. After removing IndexedGenericOp, the base class is not used anymore.

Differential Revision: https://reviews.llvm.org/D105307

[mlir][linalg][python] Add exp and log to the OpDSL.

Introduce the exp and log function in OpDSL. Add the soft plus operator to test the emitted IR in Python and C++.

Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D105420

[libcxx] [test] Fix spurious failures in the thread detach test on Windows

Make sure that the detached thread has started up before exiting
the process.

If the detached thread hasn't started up at all, and the main thread
exits, global data structures in the process are torn down, which
then can cause crashes when the thread starts up late after required
mutexes have been destroyed. (In particular, the mutex used internally
in _Init_thread_header, which is used in the initialization of
__thread_local_data()::__p, can cause crashes if the main thread already
has finished and progressed far with destruction.)

Differential Revision: https://reviews.llvm.org/D105592

[IR] Added operator delete to subclasses of User to avoid UB

Several subclasses of User override operator new without also overriding
operator delete. This means that delete expressions fall back to using
operator delete of the base class, which would be User. However, this is
only allowed if the base class has a virtual destructor which is not the
case for User, so this is UB.

See also [expr.delete] (3) for the exact wording.

This is actually detected in some cases by GCC 11's
-Wmismatched-new-delete now which is how I found this error.

Differential Revision: https://reviews.llvm.org/D103143

Revert "[mlir][MemRef] Fix DimOp folding of OffsetSizeAndStrideInterface."

This reverts commit 6c0fd4db79f2def432f761627bb8c7d4171a3237.

This simple implementation is unfortunately not extensible and needs to be reverted.
The extensible way should be to extend https://reviews.llvm.org/D104321.

[Test] Add loop deletion switch tests

Patch by Dmitry Makogon!

Differential Revision: https://reviews.llvm.org/D105543

[AArch64][SVE] Add ISel patterns for floating point compare with zero instructions

Additionally, lower the floating point compare SVE intrinsics to
SETCC_MERGE_ZERO ISD nodes to avoid duplicating ISel patterns.

Differential Revision: https://reviews.llvm.org/D105486

[mlir-reduce] Fix the memory leak and recycle unused modules.

Reviewed By: jpienaar

Differential Revision: https://reviews.llvm.org/D105416

[NFC] Mark Expected<T>::assertIsChecked() as const

Some const methods of Expected<T> invoke assertIsChecked(),
so we should mark it as const too.

Differential Revision: https://reviews.llvm.org/D105292

Support: add llvm::thread class that supports specifying stack size.

This adds a new llvm::thread class with the same interface as std::thread
except there is an extra constructor that allows us to set the new thread's
stack size. On Darwin even the default size is boosted to 8MB to match the main
thread.

It also switches all users of the older C-style `llvm_execute_on_thread` API
family over to `llvm::thread` followed by either a `detach` or `join` call and
removes the old API.

Revert "Support: add llvm::thread class that supports specifying stack size."

It's causing build failures because DefaultStackSize isn't defined everywhere
it should be and I need time to investigate.

[amdgpu] Remove the GlobalDCE pass prior to the internalization pass.

- In [D98783](https://reviews.llvm.org/D98783), an extra GlobalDCE pass
  is inserted before the internalization pass to ensure a global
  variable without users could be internalized even if there are dead
  users. Instead of inserting a dedicated optimization pass, the
  dead user checking, i.e. 'use_empty()', should be preceeded with
  constant dead user removal to ensure an accurate result.

Differential Revision: https://reviews.llvm.org/D105590

[Instcombine][NFC]Add a test for reduce+([sext/zext](<n x i1)) case, NFC.

[mlir] Fully qualify types and expressions in Interfaces

This patch adds full qualification to the types and function calls used inside of (Static)InterfaceMethod in OpInterfaces. Without this patch using many of these interfaces in a downstream project yields compiler errors as the types and default implementations are mostly copied verbatim. Without then putting using namespace mlir; in the header file of the implementations of those interfaces, compilation is impossible.

Using fully qualified lookup fixes this issue.

Differential Revision: https://reviews.llvm.org/D105619

[Internalize] Preserve variables externally initialized.

- ``externally_initialized`` variables would be initialized or modified
  elsewhere. Particularly, CUDA or HIP may have host code to initialize
  or modify ``externally_initialized`` device variables, which may not
  be explicitly referenced on the device side but may still be used
  through the host side interfaces. Not preserving them triggers the
  elimination of them in the GlobalDCE and breaks the user code.

Reviewed By: yaxunl

Differential Revision: https://reviews.llvm.org/D105135

[Instcombine]Transform reduction+(sext/zext(<n x i1>) to <n x im>) to [-]zext/trunc(ctpop(bitcast <n x i1> to in)) to im.

Some of the SPEC tests end up with reduction+(sext/zext(<n x i1>) to <n x im>) pattern, which can be transformed to [-]zext/trunc(ctpop(bitcast <n x i1> to in)) to im.
Also, reduction+(<n x i1>) can be transformed to ctpop(bitcast <n x i1> to in) & 1 != 0.

Differential Revision: https://reviews.llvm.org/D105587

Recommit: Support: add llvm::thread class that supports specifying stack size.

This adds a new llvm::thread class with the same interface as std::thread
except there is an extra constructor that allows us to set the new thread's
stack size. On Darwin even the default size is boosted to 8MB to match the main
thread.

It also switches all users of the older C-style `llvm_execute_on_thread` API
family over to `llvm::thread` followed by either a `detach` or `join` call and
removes the old API.

Moved definition of DefaultStackSize into the .cpp file to hopefully
fix the build on some (GCC-6?) machines.

[libc++] Guard testing implementation details.

The unit tests test some implementation details. As @Quuxplusone pointed
out in D96664 this should only be tested when the tests use libc++. This
addresses the issue for code already in main.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D105568

[libc++][NFC] Move monostate to its own header.

The format library uses `std::monostate`, but not a `std::variant`.
Moving `std::monostate` to its own header allows the format library to
reduce the amount of included code.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D105582

[gn build] Port 321c2ea91cb1

[DebugInfo][InstrRef][4/4] Support DBG_INSTR_REF through all backend passes

This is a cleanup patch -- we're now able to support all flavours of
variable location in instruction referencing mode. This patch updates
various tests for debug instructions to be broader: numerous code paths
try to ignore debug isntructions, and they now have to ignore the
additional DBG_PHI and DBG_INSTR_REFs that we can generate.

A small amount of rework happens for LiveDebugVariables: as we don't need
to track live intervals through regalloc any more, we can get away with
unlinking debug instructions before regalloc, then re-inserting them after.
Note that this isn't (yet) true of DBG_VALUE_LISTs, they still have to go
through live interval tracking.

In SelectionDAG, add a helper lambda that emits half-formed DBG_INSTR_REFs
for arguments in instr-ref mode, DBG_VALUE otherwise. This is one of the
final locations where DBG_VALUEs are emitted for vreg arguments.

X86InstrInfo now un-sets the debug instr number on SUB instructions that
get mutated into CMP instructions. As the instruction no longer computes a
subtraction, we can't use it for variable locations.

Differential Revision: https://reviews.llvm.org/D88898

[libc++] Workaround failures with modules on Clang ToT

[libc++][docs] Update documentation to reflect libc++'s compiler support policy

In https://lists.llvm.org/pipermail/llvm-dev/2021-March/148881.html, we
discussed updating the compiler support policy for libc++ to match more
closely what we do actually support.

This commit enshrines that policy decision in libc++'s documentation.

Differential Revision: https://reviews.llvm.org/D105563

[IR] Restore vector support for deprecated CreateGEP methods

As pointed out in post-commit review on rG8e22539067d9, it's
necessary to call getScalarType() to support GEPs with a vector
base. Dropping that call was an oversight on my side.

[libc++][ci] Stop testing on GCC previous, since we don't support it anymore

This is the first of a few commits that update the CI to match the
recently officialized compiler support policy. I'm staging those
changes to try and keep the CI green at all times, accounting how
builders refresh their Docker image.

[AMDGPU] Set LoopInfo as preserved by SIAnnotateControlFlow

The pass does not change loops, it just adds calls.

Differential Revision: https://reviews.llvm.org/D105583

[mlir][sparse] updated setter/getter comments

For the getters, it is bad practice to keep the reference
around for too long, as explained in the new comment

Reviewed By: gussmith23

Differential Revision: https://reviews.llvm.org/D105599

[LangRef] Clarify !associated

Notably, a global variable with the metadata should generally not be referenced
by a function function. E.g. -fstack-size-section usage is fine, but
-fsanitize-coverage= used to have a linker GC problem (fixed by D97430).

Reviewed By: eugenis

Differential Revision: https://reviews.llvm.org/D104933

[CostModel][X86] Account for older SSE targets with slow fp->int conversions

Both the conversion cost and the xmm->gpr transfer cost tend to be a lot higher on early SSE targets

Fix MSVC "truncation from 'int' to 'bool'" warning. NFCI.

[ScalarEvolution] Fix overflow in computeBECount.

There are two issues with the current implementation of computeBECount:

1. It doesn't account for the possibility that adding "Stride - 1" to
Delta might overflow. For almost all loops, it doesn't, but it's not
actually proven anywhere.
2. It doesn't account for the possibility that Stride is zero. If Delta
is zero, the backedge is never taken; the value of Stride isn't
relevant. To handle this, we have to make sure that the expression
returned by computeBECount evaluates to zero.

To deal with this, add two new checks:

1. Use a variety of tricks to try to prove that the addition doesn't
overflow.  If the proof is impossible, use an alternate sequence which
never overflows.
2. Use umax(Stride, 1) to handle the possibility that Stride is zero.

Differential Revision: https://reviews.llvm.org/D105216

[compiler-rt][hwasan] Define fuchsia implementations of required hwasan functions

This contains all the definitions required by hwasan for the fuchsia
implementation and can be landed independently from the remaining parts of D91466.

Differential Revision: https://reviews.llvm.org/D103936

[LangRef] Fix typo about SHF_LINK_ORDER

Mips/GlobalISel: Use correct callee calling convention

This was using the convention from the calling function.

GlobalISel: Track original argument index in ArgInfo

SelectionDAG's equivalents in ISD::InputArg/OutputArg track the
original argument index. Mips relies on this, and its currently
reinventing its own parallel CallLowering infrastructure which tracks
these indexes on the side. Add this to help move towards deleting the
custom mips handling.

Mips/GlobalISel: Remove custom splitToValueTypes

[AMDGPU] Fix indention in llc-pipeline test. NFC.

[Metadata] Decorate methods with 'const'. NFC.

- Minor coding style fix.

[mlir][Linalg] Fix tensor.extract_slice(linalg.init_tensor) canonicalization for rank-reducing extract

Differential Revision: https://reviews.llvm.org/D105636

[libc++][ci] Install Clang 11, Clang 12 and Clang ToT in the Docker image

The compiler support policy mentions that we support Clang 11 and 12, so
we should test those. We already test on Clang 12, but I'll add testers
for Clang 11 once the new Docker image is in use on all the builders.

[AMDGPU] Fix more indention in llc-pipeline test. NFC.

[clang] fix constexpr code generation for user conversions.

When building the member call to a user conversion function during an
implicit cast, the expression was not being checked for immediate
invocation, so we were never adding the ConstantExpr node to AST.

This would cause the call to the user conversion operator to be emitted
even if it was constantexpr evaluated, and this would even trip an
assert when said user conversion was declared consteval:
`Assertion failed: !cast<FunctionDecl>(GD.getDecl())->isConsteval() && "consteval function should never be emitted", file clang\lib\CodeGen\CodeGenModule.cpp, line 3530`

Fixes PR48855.

Signed-off-by: Matheus Izvekov <mizvekov@gmail.com>

Reviewed By: rsmith

Differential Revision: https://reviews.llvm.org/D105446

[X86][NFC]Add run lines for AVX512VL for masked gather test, NFC.

[CodeGen] Avoid CreateGEP with nullptr type (NFC)

In preparation for dropping support for it. I've replaced it with
a proper type where the correct type was obvious and left an
explicit getPointerElementType() where it wasn't.

[libc++] Add XFAIL for Clang ToT with modules

This is what I should have done instead of 6afd6e96ce20.

[ARM] Use matchSimpleRecurrence to simplify some code in MVEGatherScatterLowering. NFCI

Reviewed By: dmgreen

Differential Revision: https://reviews.llvm.org/D105262

[SLP][COST][X86]Improve cost model for masked gather.

Revived D101297 in its original form + added some changes in X86
legalization cehcking for masked gathers.

This solution is the most stable and the most correct one. We have to
check the legality before trying to build the masked gather in SLP.
Without this check we have incorrect cost (for SLP) in case if the masked gather
is not legal/slower than the gather. And we're missing some
vectorization opportunities.

This can be fixed in the cost model, but in this case we need to add
special checks for the cost of GEPs for ScatterVectorize node, add
special check for small trees, etc., i.e. there are a lot of corner
cases here and there, which insrease code base and make it harder to
maintain the code.

> Can't we rely on cost model to deal with this? This can be profitable for futher vectorization, when we can start from such gather loads as seed.

The question from D101297. Actually, no, it can't. Actually, simple
gather may give us better result, especially after we started
vectorization of insertelements. Plus, like I said before, the cost for
non-legal masked gathers leads to missed vectorization opportunities.

Differential Revision: https://reviews.llvm.org/D105042

[OPENMP]Do no privatize const firstprivates in target regions.

No need to emit private copyfor firstprivate constants in target
regions, we can use the original copy instead.

Differential Revision: https://reviews.llvm.org/D105647

[NVPTX] Pass explicit GEP type (NFC)

Use source element type of original GEP, as we're just changing
the address space.

[AMDGPU] Pass explicit GEP type in printf transform (NFC)

This code is working on an i8*. Avoid nullptr element type in
preparation for removing support.

[AMDGPU] Simplify GEP construction (NFC)

Noticed while making a related change. This code was doing
something really peculiar: Creating an APInt by parsing a string.
And then creating a SmallVector with one element to create the
GEP.

Instead create the APInt from integers and directly pass the single
index to GetElementPtrInst::Create().

[CodeGen] Avoid nullptr arg to CreateStructGEP (NFC)

For now just make the getPointerElementType() explicit.

[SLP]Improve vectorization of stores.

Patch tries to improve the vectorization of stores. Originally, we just
check the type and the base pointer of the store.
Patch adds some extra checks to avoid non-profitable vectorization
cases. It includes analysis of the scalar values to be stored and
triggers the vectorization attempt only if the scalar values have
same/alt opcode and are from same basic block, i.e. we don't end up
immediately with the gather node, which is not profitable.
This also improves compile time by filtering out non-profitable cases.

Part of D57059.

Differential Revision: https://reviews.llvm.org/D104122

[msan] Add funel shift tests

For https://bugs.llvm.org/show_bug.cgi?id=50840

[msan] Handle funnel shifts

Fixes https://bugs.llvm.org/show_bug.cgi?id=50840

Differential Revision: https://reviews.llvm.org/D105387

[GWP-ASan] Change unreachable -> trap to work around DCE bug.

trapOnAddress is designed to SEGV on a specific address. Unfortunately,
with an IR change, __builtin_unreachable() ends up doing DCE on things
that have side effects, like the load that causes the trap.

Change to __builtin_trap() to avoid the optimisation.

Root cause is still an LLVM bug, and tracked in
https://bugs.llvm.org/show_bug.cgi?id=47480.

Reviewed By: eugenis

Differential Revision: https://reviews.llvm.org/D105654

[NFC][compiler-rt][fuchsia] Add InitShadowBounds declaration to header

Forgot to include this as a part of
a11aea68a4b31e72bc0b84bde5f3210048287d28.

PR51018: Remove explicit conversions from SmallString to StringRef to future-proof against C++23

C++23 will make these conversions ambiguous - so fix them to make the
codebase forward-compatible with C++23 (& a follow-up change I've made
will make this ambiguous/invalid even in <C++23 so we don't regress
this & it generally improves the code anyway)

PR51018: Disallow explicit construction of StringRef from SmallString due to ambiguity in C++23

See bug for full details, but basically there's an upcoming ambiguity in
the conversion in `StringRef(SomeSmallString)` - either the implicit
conversion operator (SmallString::operator StringRef) could be used, or
the std::string_view range-based ctor (& then `StringRef(std::string_view)`
would be used)

To address this, make such a conversion invalid up-front - most uses are
more tersely written as `SomeSmallString.str()` anyway, or more clearly
written as `StringRef x = y;` rather than `StringRef x(y);` - so if you
hit this in out-of-tree code, please update in one of those ways.
Hopefully I've fixed everything in tree prior to this patch landing.

Revert "PR51018: Disallow explicit construction of StringRef from SmallString due to ambiguity in C++23"

This reverts commit e2d30846327c7ec5cc9d2a46aa9bcd9c2c4eff93.

MSVC doesn't seem to resolve the intended ambiguity in implicit
conversion contexts correctly: https://godbolt.org/z/ee16aqv4v

[mlir][bazel] Added missing MathDialect dep to LinalgOps target

LinalgOps needs MathDialect as a dependency for it to build correctly.

Differential Revision: https://reviews.llvm.org/D105656

[InstCombine][NFC]Use only `replaceInstUsesWith`, NFC.

[ARM] Extra v8i16 -> i64 reduction tests with loads. NFC

[WebAssembly] Scalarize extract_vector_elt of binops

Override the `shouldScalarizeBinop` target lowering hook using the same
implementation used in the x86 backend. This causes `extract_vector_elt`s of
vector binary ops to be scalarized if the scalarized version would be supported.

Differential Revision: https://reviews.llvm.org/D105646

Mark TensorDialect legal and PadTensor op illegal

`GeneralizePadTensorOpPattern` might generate `tensor.dim` op so the
TensorDialect should be marked legal. This pattern should also
transform all `linalg.pad_tensor` ops so mark those as illegal. Those
changes are missed from a previous change in
https://reviews.llvm.org/D105293

Reviewed By: silvas

Differential Revision: https://reviews.llvm.org/D105642

[ARM] Pass 2 instead of 0 to PHINode::Create in MVEGatherScatterLowering. NFC

This parameter controls how much space is reserved for incoming
values. There are always going to be 2 incoming values in this case.

While there remove the unused std::vector right below.

Found while looking at porting this code to RISCV.

[AMDGPU] Mark more SOP instructions as rematerializable

The rest of the SOP instructions implicitly set SCC and not
suitable for the rematerialization.

Differential Revision: https://reviews.llvm.org/D105670

[mlir][sparse] Add Merger unit tests

We opt to use unit tests rather than check tests as the lattice/merger code is a small C++ component with a well-defined API. Testing this API via check tests would be far less direct and readable. In addition, as the check tests will only be able to test the API indirectly, the tests may break based on unrelated changes; e.g. changes in linalg.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D104956

Add AddDiscriminatorsPass to NPM default O0 pipeline

AddDiscriminatorsPass is in Legacy PM's O0 pipeline. This patch did the same
for NPM O0 pipeline.

Reviewed By: aeubanks, MaskRay

Differential Revision: https://reviews.llvm.org/D105650

Add documentation for -fsanitize-address-use-after-return.

for issue: https://github.com/google/sanitizers/issues/1394

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D104155

[mlir-reduce] Fix the grammer in the doc

Reviewed By: jpienaar

Differential Revision: https://reviews.llvm.org/D105623

[NFC][ScalarEvolution] Cleanup howManyLessThans.

In preparation for D104075. Some NFC cleanup, and some test coverage for
planned changes.

Revert "[mlir][sparse] Add Merger unit tests"

This reverts commit 0d0cff3ace39378acfc66d6564dc99e19b8a561f.

The build is broken with GCC 5.4

[RISCV][test] Add new tests for mul optimization in the zba extension with SH*ADD

This patch will show the following optimization by future patches.

(mul x imm) -> (SH1ADD x, (SLLI x, bits)) when imm = 2^n + 2.
(mul x imm) -> (SH2ADD x, (SLLI x, bits)) when imm = 2^n + 4.
(mul x imm) -> (SH3ADD x, (SLLI x, bits)) when imm = 2^n + 8.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D105614

[ASan][AMDGPU] Make shadow offset match X86 on Linux

This patch explicitly sets the shadow offset for
AMDGPU to match that of X86 on Linux.

Reviewed By: vitalybuka

https://reviews.llvm.org/D105282

Revert "GlobalISel/AArch64: don't optimize away redundant branches at -O0"

This reverts commit 458c230b5ef893238d2471fcff27cd275e8026d5.

This broke LLDB buildbot testcase where breakpoint set at start of loop
failed to hit. https://lab.llvm.org/buildbot/#/builders/96/builds/9404

https://github.com/llvm/llvm-project/blob/main/lldb/test/API/commands/process/attach/main.cpp#L15

Differential Revision: https://reviews.llvm.org/D105238

Update Bazel overlay in GPUToGPURuntimeTransforms.

[PowerPC] Remove implicit use register after transformToImmFormFedByLI()

When the instruction has imm form and fed by LI, we can remove the redundat LI instruction.
Below is an example:
```
    renamable $x5 = LI8 2
    renamable $x4 = exact SRD killed renamable $x4, killed renamable $r5, implicit $x5
```

will be converted to:
```
   renamable $x5 = LI8 2
   renamable $x4 = exact RLDICL killed renamable $x4, 62, 2,  implicit killed $x5
```

But when we do this optimization, we forget to remove implicit killed $x5
This bug has caused a lnt case error. This patch is to fix above bug.

Reviewed By: #powerpc, shchenz

Differential Revision: https://reviews.llvm.org/D85288

[CMake] Disable -fno-semantic-interposition for GCC<10.3 on SystemZ

`-fno-semantic-interposition` was added for GCC in D102453, but some MLIR tests
on SystemZ failed with GCC<10.3 due to a bug.

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D105453

[RISCV][clang] Add macro __riscv_zvlsseg for RVV Zvlsseg builtins

Add extension macro __riscv_zvlsseg to enable Zvlsseg builtins only
with target feature Zvlsseg.

Reviewed By: HsiangKai

Differential Revision: https://reviews.llvm.org/D105626