And as being reported by Florian Hahn, there's a hit
in MultiSource/Benchmarks/mafft from the test-suite on X86 with -O3 -flto,
so reverting until addressed.

This reverts commit 71e0b82c.

Replace isBuildVectorAllZeros check by using the Zeroable bitmask instead.

It's good to reuse the DWARF parser in lib/DebugInfo so that we don't
need to maintain a separate parser in client side (obj2yaml). Besides,
A test case is added whose length field is a very huge value which makes
obj2yaml stuck when parsing the section.

Reviewed By: jhenderson

Differential Revision: https://reviews.llvm.org/D86192

Address some post-commit feedback on D85968.

TypeSystemClang::CreateTypedef was creating a typedef in the right
DeclContext, but it was not actually adding it as a child of the
context. The resulting inconsistent state meant that we would be unable
to reference the typedef from an expression directly, but we could use
them if they end up being pulled in by some previous subexpression
(because the ASTImporter will set up the correct links in the expression
ast).

This patch adds the typedef to the decl context it is created in.

Differential Revision: https://reviews.llvm.org/D86140

canonicalizeShuffleMaskWithCommute should have already ensured the lower elements are from V1, we do have test coverage for this already.

Thanks to @fhahn for the test case.

I missed updating these with:
rG23bd33c6

It's been a month since 2f3862eb,
and no new bug reports about the threshold were filled,
so let's bump it again and wait again.

This adds a minor test case fix to previously submitted AArch64 SVE
ptrace support. This was failing on LLDB/AArch64 Linux buildbot.

Differential Revision: https://reviews.llvm.org/D79699

[X86][AVX] computeKnownBitsForTargetNode - add VTRUNC/VTRUNCS/VTRUNCUS known zero upper elements handling.

Like many of the AVX512 conversion ops, the VTRUNC ops guarantee the upper destination elements are zero.

This patch adds NativeRegisterContext_arm64 ptrace routines to access
AArch64 SVE register set. This patch also adds a test-case to test
AArch64 SVE register access and dynamic size configuration capability.

Reviewed By: labath

Differential Revision: https://reviews.llvm.org/D79699

Add handling for storing the extracted lower (truncated bits) element from a X86ISD::VTRUNC node - this can be lowered to a generic truncated store directly.

Differential Revision: https://reviews.llvm.org/D86158

This function returns a struct `BuiltinVectorTypeInfo` that contains
the builtin vector's element type, element count and number of vectors
(used for vector tuples).

Reviewed By: c-rhodes

Differential Revision: https://reviews.llvm.org/D86100

Introduce two new AAs. AAICVTrackerFunctionReturned which checks if a
function can have a unique ICV value after it is finished, and
AAICVCallSiteReturned which checks AAICVTrackerFunctionReturned for a
call site. This enables us to check the value of a call and if it
changes the ICV. This also changes the approach in
`getReplacementValues()` to a worklist-based approach so we can explore
all relevant BBs.

Differential Revision: https://reviews.llvm.org/D85544

Intrinsic properties can now be set to default and applied to all
intrinsics. If the attributes are not needed, the user can opt-out by
setting the DisableDefaultAttributes flag to true.

Differential Revision: https://reviews.llvm.org/D70365

Used InstCombine to enable VMLAV and Add instructions to generate VMLAVA instead with tests.

This implements the assemble and disassemble support of RISCV Vector
extension Zvlsseg instructions, base on the 0.9 spec version.

Reviewed  by HsiangKai

Differential Revision: https://reviews.llvm.org/D84416

Regarding this bug in Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=46451

I went ahead and fixed the regexp pattern and now Python script is able
to extract vplan graphs from the log files. Additionally some test for
this would be nice to have but I'm not sure are Python scripts tested
in LLVM and if so where they live.

Reviewed By: fhahn

Differential Revision: https://reviews.llvm.org/D86068

This patch fixes a bug which skipped
adding predicate matcher for a pattern in many cases.
For example, if predicate is Load and
its memoryVT is non-null then the loop
continues and never reaches to the end which
adds the predicate matcher. This patch moves the
matcher addition to the top of the loop
so that it gets added regardless of contextual checks
later in the loop.
Other way to fix this issue is to remove all "continue" statements
in checks and let the loop continue till end.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D83034

Legacy DSE uses NumRedundantStores, while MemorySSA DSE uses
NumNoopStores. We should just use the same counter.

This reverts commit cacfb02d.

Reverting due to buildbot failures.

In our discussion D79699 SVE ptrace register access support we decide to
invalidate register context cached data on every stop instead of doing
at before Step/Resume.

InvalidateAllRegisters was added to facilitate flushing of SVE register
context configuration and cached register values. It now makes more
sense to move invalidation after every stop where we initiate SVE
configuration update if needed by calling ConfigureRegisterContext.

Reviewed By: labath

Differential Revision: https://reviews.llvm.org/D84501

This patch updates LLDB's in house version of SVE ptrace/sig macros by
converting them into constants and inlines. They are housed under sve
namespace and are used by process elf-core for reading SVE register data.

Reviewed By: labath

Differential Revision: https://reviews.llvm.org/D85641

Differential Revision: https://reviews.llvm.org/D86130

In DAGTypeLegalizer::GenWidenVectorLoads the algorithm assumes it only
ever deals with fixed width types, hence the offsets for each individual
store never take 'vscale' into account. I've changed the code in that
function to use TypeSize instead of unsigned for tracking the remaining
load amount. In addition, I've changed the load loop to use the new
IncrementPointer helper function for updating the addresses in each
iteration, since this handles scalable vector types.

Also, I've added report_fatal_errors in GenWidenVectorExtLoads,
TargetLowering::scalarizeVectorLoad and TargetLowering::scalarizeVectorStores,
since these functions currently use a sequence of element-by-element
scalar loads/stores. In a similar vein, I've also added a fatal error
report in FindMemType for the case when we decide to return the element
type for a scalable vector type.

I've added new tests in

  CodeGen/AArch64/sve-split-load.ll
  CodeGen/AArch64/sve-ld-addressing-mode-reg-imm.ll

for the changes in GenWidenVectorLoads.

Differential Revision: https://reviews.llvm.org/D85909

mtune was previously ignored by the compiler so I'm not sure this
did anything. But after D85384 we're starting to support mtune
and this code is now causing a couple test failures on MacOS.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D86128

Some build bot has lld in the directory name, which caused pattern match
issue in the list test.