Changes

[AMDGPU] Fix gcc warnings about unused variables [NFC]

[mlir] Fix build on Windows on Arm

clang-cl errors out while handling the templated version of tgfmt. This
patch works around the issue by explicitly choosing the non-templated
version of tgfmt, which takes an ArrayRef<std::string>.

More details in this thread:
https://lists.llvm.org/pipermail/cfe-dev/2021-September/068936.html

Thanks @Mehdi Amini for suggesting the fix :)

Differential Revision: https://reviews.llvm.org/D110223

[clang][tooling] NFC: Refactor command-line diagnostic tests

This patch uses a different command-line arguments to test `clang::tooling::ToolInvocation` that are not specific to Darwin.

Reviewed By: dexonsmith

Differential Revision: https://reviews.llvm.org/D110160

[X86][FP16] Combine the FADD(A, FMA(B, C, 0)) to FMA(B, C, A)

This patch is to support transform something like
_mm512_add_ph(acc, _mm512_fmadd_pch(a, b, _mm512_setzero_ph()))
to _mm512_fmadd_pch(a, b, acc).

Differential Revision: https://reviews.llvm.org/D109953

Wrap xar/xar.h include in extern "C" block

Without such wrapping, linking lld fails with missing symbols because of
C++ symbol mangling with older versions of the MacOSX SDK, in which
xar.h doesn't have an extern "C" block itself.

Reviewed By: #lld-macho, thakis

Differential Revision: https://reviews.llvm.org/D110224

[NFC] Wrap MLIR addAffineForOpDomain warning with LLVM_DEBUG

Current warning message in method `addAffineForOpDomain` of mlir/lib/Analysis/AffineStructures.cpp is being printed to the stdout/stderr.
This patch redirects the warning with LLVM_DEBUG following standard llvm practice.

Reviewed By: bondhugula

Differential Revision: https://reviews.llvm.org/D108340

[clangd] Ensure lambda init-capture gets semantic token

Prior to this patch, CollectExtraHighlightings would incorrectly produce
a token for the init-capture's type which overlapped the name and
resulted in both being dropped.

Fixes https://github.com/clangd/clangd/issues/868

Differential Revision: https://reviews.llvm.org/D110130

[ModuleInlinerWrapperPass] Do some naive printing of wrapped pipeline with -print-pipeline-passes

Bisecting and reducing opt pipelines that includes the
ModuleInlinerWrapperPass has turned out to be a bit problematic.
This is far from perfect (it still lacks information about inline
advisor params etc.), but it should give some kind of hint to what
the wrapped pipeline looks like when using -print-pipeline-passes.

Reviewed By: aeubanks, mtrofin

Differential Revision: https://reviews.llvm.org/D109878

[DAGCombiner] Rename isAlias as mayAlias. NFC

Differential Revision: https://reviews.llvm.org/D110062

[TII] Remove the MFI argument to convertToThreeAddress. NFC.

This simplifies the API and addresses a FIXME in
TwoAddressInstructionPass::convertInstTo3Addr.

Differential Revision: https://reviews.llvm.org/D110229

[Driver] Correctly handle static C++ standard library

When statically linking C++ standard library, we shouldn't add -Bdynamic
after including the library on the link line because that might override
user settings like -static and -static-pie. Rather, we should surround
the library with --push-state/--pop-state to make sure that -Bstatic
only applies to C++ standard library and nothing else. This has been
supported since GNU ld 2.25 (2014) so backwards compatibility should
no longer be a concern.

Differential Revision: https://reviews.llvm.org/D110128

Revert "[Driver] Correctly handle static C++ standard library"

This reverts commit 5e28c892d06f95600f8b6290ad4de38bfe142637 as
the linker on the clang-ppc64le-rhel bot doesn't seem to support
--push-state/--pop-state.

[libc] Add optimized memset for AArch64

Differential Revision: https://reviews.llvm.org/D107848

[InstSimplify] Don't lose inbounds when simplifying a GEP

I noticed this while working on a (ptrtoint (gep null, x)) -> x fold.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D110168

tsan: remove unnecessary enum values (NFC)

Remove unnecessary enum values in the memory profiler.
There is no point in spelling them, it can only lead to bugs
and larger diffs when values are added/removed.

Reviewed By: melver

Differential Revision: https://reviews.llvm.org/D110263

tsan: remove DontDumpShadow from Go build

DontDumpShadow is used only in InitializeShadowMemory which is Go-only.

Depends on D110263.

Reviewed By: melver

Differential Revision: https://reviews.llvm.org/D110264

tsan: add another deep stack test

Add a test for a trace corner case that lead to a bug
in experimental runtime replacement.
Since it passes with the current runtime it makes sense
to submit it on its own.

Depends on D110264.

Reviewed By: melver

Differential Revision: https://reviews.llvm.org/D110265

tsan: use RawShadow instead of u64 more

Fix few remaining cases where we use u64 instead of the new RawShadow type.

Depends on D110265.

Reviewed By: melver

Differential Revision: https://reviews.llvm.org/D110266

tsan: move shadow declaration into a separate header file (NFC)

tsan_rtl.h is very huge and contains too many things.
Move FastState and Shadow types into a new tsan_shadow.h file.
This also allows to use FastState/Shadow in other header files
without creating circular dependencies (which most likely will
happen today).

Reviewed By: melver

Differential Revision: https://reviews.llvm.org/D110272

[fir][NFC] inline trival isa_<type> functions

This patch is part of the upstreaming effort from fir-dev branch and sync changes. Inline trival `isa_<type>` functions.

Co-authored-by: schweitzpgi

Reviewed By: kiranchandramohan

Differential Revision: https://reviews.llvm.org/D110233

[RISCV][VP] Add support for VP_REDUCE_* operations

This patch adds codegen support for lowering the vector-predicated
reduction intrinsics to RVV instructions. The process is similar to that
of the other reduction intrinsics, save for the fact that every VP
reduction has a start value. We reuse the existing custom "VL" nodes,
adding extra patterns where required to handle non-true masks.

To support these nodes, the `RISCVISD::VECREDUCE_*_VL` nodes have been
given an explicit "merge" operand. This is to faciliate the VP
reductions, where we must be careful to ensure that even if no operation
is performed (when VL=0) we still produce the start value. The RVV
reductions don't update the destination register under these conditions,
so we tie the splatted start value to the output register.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D107657

[lldb] Fix DomainSocket::GetSocketName for unnamed sockets

getpeername will return addrlen = 2 (sizeof sa_family_t) for unnamed
sockets (those not assigned a name with bind(2)). This is typically true
for client sockets as well as those created by socketpair(2).

This GetSocketName used to crash for sockets which were connected to
these kinds of sockets. Now it returns an empty string.

[BasicBlockUtils] Fixup of an assumed typo in MergeBlockIntoPredecessor

The NFC commit e5692a564a73ef63b7b changed the logic for
DomTreeUpdates to use the range [succ_begin, succ_begin) when
looking for SuccsOfPredBB rather than using [succ_begin, succ_end).

As the commit was NFC this is identified as a typo (it has been
discussed briefly in phabricator).

The typo was found when inspecting the code, so I've got no idea if
changing back to the old range has any significant impact (such as
solving any PR:s or causing some new problems). But at least this
restores the code to the originally indented behavior.

[CodeGen] RegisterCoalescer::buildVRegToDbgValueMap - use const-ref value in for-range loop. NFCI.

Avoid unnecessary copies, reported by MSVC static analyzer.

[CodeGen] ProcessSDDbgValues - use const-ref value in for-range loop. NFCI.

Avoid unnecessary copies, reported by MSVC static analyzer.

[RISCV] Add missing op type OPERAND_UIMM2, OPERAND_UIMM3 and OPERAND_UIMM7 for verifyInstruction

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D110307

[DSE] Track earliest escape, use for loads in isReadClobber.

At the moment, DSE only considers whether a pointer may be captured at
all in a function. This leads to cases where we fail to remove stores to
local objects because we do not check if they escape before potential
read-clobbers or after.

Doing context-sensitive escape queries in isReadClobber has been removed
a while ago in d1a1cce5b130 to save compile-time. See PR50220 for more
context.

This patch introduces a new capture tracker, which keeps track of the
'earliest' capture. An instruction A is considered earlier than instruction
B, if A dominates B. If 2 escapes do not dominate each other, the
terminator of the common dominator is chosen. If not all uses cannot be
analyzed, the earliest escape is set to the first instruction in the
function entry block.

If the query instruction dominates the earliest escape and is not in a
cycle, then pointer does not escape before the query instruction.

This patch uses this information when checking if a load of a loaded
underlying object may alias a write to a stack object. If the stack
object does not escape before the load, they do not alias.

I will share a follow-up patch to also use the information for call
instructions to fix PR50220.

In terms of compile-time, the impact is low in general,
    NewPM-O3: +0.05%
    NewPM-ReleaseThinLTO: +0.05%
    NewPM-ReleaseLTO-g: +0.03

with the largest change being tramp3d-v4 (+0.30%)
http://llvm-compile-time-tracker.com/compare.php?from=1a3b3301d7aa9ab25a8bdf045c77298b087e3930&to=bc6c6899cae757c3480f4ad4874a76fc1eafb0be&stat=instructions

Compared to always computing the capture information on demand, we get
the following benefits from the caching:
NewPM-O3: -0.03%
NewPM-ReleaseThinLTO: -0.08%
NewPM-ReleaseLTO-g: -0.04%

The biggest speedup is tramp3d-v4 (-0.21%).
http://llvm-compile-time-tracker.com/compare.php?from=0b0c99177d1511469c633282ef67f20c851f58b1&to=bc6c6899cae757c3480f4ad4874a76fc1eafb0be&stat=instructions

Overall there is a small, but noticeable benefit from caching. I am not
entirely sure if the speedups warrant the extra complexity of caching.
The way the caching works also means that we might miss a few cases, as
it is less precise. Also, there may be a better way to cache things.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D109844

tsan: remove expected race leftover

Remove nmissed_expected variable.
It's a leftover from removed "expected race" feature and is never incremented.

Reviewed By: melver

Differential Revision: https://reviews.llvm.org/D110321

[libc++] Remove uses of _LIBCPP_HAS_NO_VARIABLE_TEMPLATES

All supported compilers provide support for variable templates now.

Differential Revision: https://reviews.llvm.org/D110284

[CodeGen] update test file to not run the entire LLVM optimizer; NFC

Clang regression tests should not break when changes are made to
the LLVM optimizer. This file broke on the 1st attempt at D110170,
so I'm trying to prevent that on another try.

Similar to other files in this directory, we make a compromise and
run -mem2reg to reduce noise by about 1000 lines out of 5000+ CHECK lines.

[gn build] (semi-manually) port 702cb7afe9de

[gn build] (manually) port ac191bcc99e2f

[gn build] Port f4abdb0c074b

[InstCombine] fold cast of right-shift if high bits are not demanded (2nd try)

The 1st try at this was reverted because it caused an infinite loop in instcombine.
That should be fixed after:
1cd6b44f267b

(masked) trunc (lshr X, C) --> (masked) lshr (trunc X), C

Narrowing the shift should be better for analysis and can lead
to follow-on transforms as shown.

Attempt at a general proof in Alive2:
https://alive2.llvm.org/ce/z/tRnnSF

Here are a couple of the specific tests:
https://alive2.llvm.org/ce/z/bCnTp-
https://alive2.llvm.org/ce/z/TfaHnb

Differential Revision: https://reviews.llvm.org/D110170

[CostModel][X86] Increase i64 mul cost from 1 to 2

Only the most recent cpus support really 1cy 64-bit multiplies, and the X64 cost table represents a realistic worst case. The 1cy value was also discouraging vectorization when most vXi64 PMULDQ expansions aren't actually slower than scalarization.

Noticed while investigating PR51436.

clangd: Do not report inline overrides twice

... in textDocument/references.

Reviewed By: sammccall

Differential Revision: https://reviews.llvm.org/D110324

[ORC][examples] Fix segfault in LLJITWithTargetProcessControl

Issue reported during review D110260

[Sema] Fix a null pointer reference crash.

Differential Revision: https://reviews.llvm.org/D110315

[AMDGPU] Avoid processing functions in amdgpu-propagate-attributes pass for shaders

The pass amdgpu-propagate-attributes ("Early/Late propagate attributes
from kernels to functions") is currently run also for shaders, where
it does nothing. Modify the check so the pass only processes functions
for kernels.

Differential Revision: https://reviews.llvm.org/D109961

Pass some DataLayout arguments by const-ref

Avoid unnecessary copies, reported by MSVC static analyzer.

Fix buildbot error.

-frounding-math is not available for all targets.

Support: Fix missing whitespace in comment for inconvertibleErrorCode(), NFC

[x86] add AVX run to tests of fcmp logic; NFC

The ISA before AVX has predicate gaps for both fcmp
codegen alternatives, so that requires a more
complicated fix to get ideal asm in all cases.

[lldb] [gdb-remote] Refactor getting remote regs to use local vector

Refactor remote register getters to collect them into a local
std::vector rather than adding them straight into DynamicRegisterInfo.
The purpose of this change is to lay groundwork for switching value_regs
and invalidate_regs to use local LLDB register numbers rather than
remote numbers.

Differential Revision: https://reviews.llvm.org/D110025

[lldb] [gdb-remote] Use local regnos for value_regs/invalidate_regs

Switch the gdb-remote client logic to use local (LLDB) register numbers
in value_regs/invalidate_regs rather than remote regnos. This involves
translating regnos received from lldb-server.

Differential Revision: https://reviews.llvm.org/D110027

[clang] Use portable "#!/usr/bin/env bash" shebang for tools and utils.

Reviewed By: JDevlieghere

Differential Revision: https://reviews.llvm.org/D110041

[RegAlloc] Cast uint8_t to unsigned before printing it.

raw_ostream interprets uint8_t as wanting to print a character
with that ASCII value. In this case the uint8_t is an integer
that we want to print.

[SampleFDO] Remove redundant declarations (NFC)

Note that these are declared in SampleProfileLoaderBaseUtil.h, which
SampleProfileLoaderBaseImpl.h includes.

[PowerPC] Add range checks for P10 Vector Builtins

This patch adds range checking for some Power10 altivec builtins and
changes the signature of a builtin to match documentation. For `vec_cntm`,
range checking is done via SemaChecking. For `vec_splati_ins`, the second
argument is masked to extract the 0th bit so that we always receive either a `0`
or a `1`.

Reviewed By: lei, amyk

Differential Revision: https://reviews.llvm.org/D109710

[libc++] Use CMake interface targets to setup benchmark flags

This is a re-application of da0592e4c8df which was reverted in
1454018dc1d9 because it was incompatible with older CMakes.
Instead, disable the benchmarks when CMake is too old to
support those idioms.

Differential Revision: https://reviews.llvm.org/D110285

[LiveIntervals] Fix repairOldRegInRange for simple def cases

The fix applied in D23303 "LiveIntervalAnalysis: fix a crash in repairOldRegInRange"
was over-zealous. It would bail out when the end of the range to be
repaired was in the middle of the first segment of the live range of
Reg, which was always the case when the range contained a single def of
Reg.

This patch fixes it as suggested by Matthias Braun in post-commit review
on the original patch, and tests it by adding -early-live-intervals to
a selection of existing lit tests that now pass.

(Note that D23303 was originally applied to fix a crash in
SILoadStoreOptimizer, but that is now moot since D23814 updated
SILoadStoreOptimizer to run before scheduling so it no longer has to
update live intervals.)

Differential Revision: https://reviews.llvm.org/D110238

Revert "[lldb] [gdb-remote] Use local regnos for value_regs/invalidate_regs"

This reverts commit 6fbed33d4a7de2229c40e6318f223092d3a23848.
The prerequisite commit is causing regressions.

Revert "[lldb] [gdb-remote] Refactor getting remote regs to use local vector"

This reverts commit b03e701c145365ba339657ead54a2e0cc5c02776.  This is
causing regressions when XML support is disabled.

Resolve {GlobalValue,GloalIndirectSymol}::getBaseObject confusion

While both GlobalAlias and GlobalIFunc are GlobalIndirectSymbol, their
`getIndirectSymbol()` usage is quite different (GlobalIFunc's resolver
is an entity different from GlobalIFunc itself).

As discussed on https://lists.llvm.org/pipermail/llvm-dev/2020-September/144904.html
("[IR] Modelling of GlobalIFunc"), the name `getBaseObject` is confusing when
used with GlobalIFunc.

To resolve the confusion:

* Move GloalIndirectSymol::getBaseObject to GlobalAlias:: (GlobalIFunc should use `getResolver` instead)
* Change GlobalValue::getBaseObject not to inspect GlobalIFunc. Note: the function has 7 references.
* Add GlobalIFunc::getResolverFunction to peel off potential ConstantExpr indirection
  (`strlen` in `test/LTO/Resolution/X86/ifunc.ll`)

Note: GlobalIFunc::getResolver (like GlobalAlias::getAliasee which does not peel
off ConstantExpr indirection) is kept to be used by ValueEnumerator.

Reviewed By: ibookstein

Differential Revision: https://reviews.llvm.org/D109792

[JumpThreading] Ignore free instructions

This is basically D108837 but for jump threading. Free instructions
should be ignored for the threading decision. JumpThreading already
skips some free instructions (like pointer bitcasts), but does not
skip various free intrinsics -- in fact, it currently gives them a
fairly large cost of 2.

Differential Revision: https://reviews.llvm.org/D110290

[fir][NFC] rename canBePointerOrHeapElementType to cannotBePointerOrHeapElementType

This patch is part of the upstreaming effort from fir-dev branch.
Rename the function so the name conveys better what it does.

Reviewed By: jeanPerier

Differential Revision: https://reviews.llvm.org/D110323

Co-authored-by: schweitz
Co-authored-by: jeanPerier

Revert "[LiveIntervals] Fix repairOldRegInRange for simple def cases"

This reverts commit 8229cb74125322ff337cfe316ab35c6ebf412bde.

It was failing on buildbots with expensive checks enabled.

[fir][NFC] Remove fir.cmpf replaced by mlir.cmpf

fir.cmpf op is not necessary anymore as it is replaced by mlir.cmpf.
This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: jeanPerier

Differential Revision: https://reviews.llvm.org/D110327

Co-authored-by: schweitzpgi
Co-authored-by: jeanPerier

[ARM] Extra tests for unpredicated qr MVE intrinsics.

[libc++] Remove unused macro in __config

That macro was being defined but not used anywhere in libc++, so it
must be safe to remove it.

As a fly-by fix, also remove mentions of this macro in other places
in LLVM, to make sure they were not depending on the value defined in
libc++.

Differential Revision: https://reviews.llvm.org/D110289

[flang] Implement READ(SIZE=) and INQUIRE(IOLENGTH=) in runtime

Count input characters corresponding to formatted edit descriptors
for READ(SIZE=); count output bytes for INQUIRE(IOLENGTH=).

The I/O APIs GetSize() and GetLength() were adjusted to return
std::size_t as function results.

Basic unit tests were added (and others fixed).

Differential Revision: https://reviews.llvm.org/D110291

[compiler-rt][profile] Add padding after binary IDs

Some tests with binary IDs would fail with error: no profile can be merged.
This is because raw profiles could have unaligned headers when emitting binary
IDs. This means padding should be emitted after binary IDs are emitted to
ensure everything else is aligned. This patch accounts for that padding in
__llvm_write_binary_ids.

Differential Revision: https://reviews.llvm.org/D110188

[lldb] Add a C language REPL to test LLDB's REPL infrastructure

LLDB has a bunch of code that implements REPL support, but all that code is
unreachable as no language in master currently has an implemented REPL backend.
The only REPL that exists is in the downstream Swift fork. All patches for this
generic REPL code therefore also only have tests downstream which is clearly not
a good situation.

This patch implements a basic C language REPL on top of LLDB's REPL framework.
Beside implementing the REPL interface and hooking it up into the plugin
manager, the only other small part of this patch is making the `--language` flag
of the expression command compatible with the `--repl` flag. The `--repl` flag
uses the value of `--language` to see which REPL should be started, but right
now the `--language` flag is only available in OptionGroups 1 and 2, but not in
OptionGroup 3 where the `--repl` flag is declared.

The REPL currently can currently only start if a running target exists. I'll add
the 'create and run a dummy executable' logic from Swift (which is requires when
doing `lldb --repl`) when I have time to translate all this logic to something
that will work with Clang.

I should point out that the REPL currently uses the C expression parser's
approach to persistent variables where only result variables and the ones
starting with a '$' are transferred between expressions. I'll fix that in a
follow up patch. Also the REPL currently doesn't work in a non-interactive
terminal. This seems to be fixed in the Swift fork, so I assume one of our many
REPL downstream changes addresses the issue.

Reviewed By: JDevlieghere

Differential Revision: https://reviews.llvm.org/D87281

[WebAssembly] Add prototype relaxed SIMD fma/fms instructions

Add experimental clang builtins, LLVM intrinsics, and backend definitions for
the new {f32x4,f64x2}.{fma,fms} instructions in the relaxed SIMD proposal:
https://github.com/WebAssembly/relaxed-simd/blob/main/proposals/relaxed-simd/Overview.md.
Do not allow these instructions to be selected without explicit user opt-in.

Differential Revision: https://reviews.llvm.org/D110295

[lldb] [gdb-remote] Refactor getting remote regs to use local vector

Refactor remote register getters to collect them into a local
std::vector rather than adding them straight into DynamicRegisterInfo.
The purpose of this change is to lay groundwork for switching value_regs
and invalidate_regs to use local LLDB register numbers rather than
remote numbers.

Differential Revision: https://reviews.llvm.org/D110025

[lldb] [gdb-remote] Use local regnos for value_regs/invalidate_regs

Switch the gdb-remote client logic to use local (LLDB) register numbers
in value_regs/invalidate_regs rather than remote regnos. This involves
translating regnos received from lldb-server.

Differential Revision: https://reviews.llvm.org/D110027

Revert "Diagnose -Wunused-value based on CFG reachability"

This reverts commit cbbf2e8c8ae7730ff0121f4868de4a7d188feb65.
It seems causing diagnoses in SFINAE context.

[x86] move combiner state check into convertIntLogicToFPLogic(); NFC

This function can be adapted to solve bugs like PR51245,
but it could require differentiating the combiner timing
between the existing and new transforms.

[x86] add AVX512 run for fcmp+logic ops; NFC

Suggested in D110342

[RISCV] Add more tests for (and (srl x, C2), C1) that can be improved by using a pair of shifts. NFC

These tests have C1 as a shifted mask having C2 leading zeros and some
number of trailing zeros, C3. We can select this as
(slli (srli x, C2+C3), C3) or (slli (srliw x, C2+C3), C3).

[RISCV] Add more isel optimizations for (and (shr x, c2), c1).

Turn (and (shr x, c2), c1) -> (slli (srli x, c2+c3), c3) if c1 is a
shifted mask with c2 leading zeros and c3 trailing zeros.

When the leading zeros is C2+32 we can use SRLIW in place of SRLI.

[AArch64] Update some sve-fixed-length test checks.

Some of these test show very poor code generation. Updating the tests
to make the tests more maintainable and prevent problems from being
hidden behind badly written test checks. Also in some of them the check
lines were using incorrect prefixes.

These are not-quite auto-generated. They are generated with the normal
update scripts and then uninteresting checks are removed, which at least
makes the test _more_ maintainable without materially changing what they
are testing.

I have otherwise attempted to not alter what is tested.

[clang-format] Add Left/Right Const fixer capability

Developers these days seem to argue over east vs west const like they used to argue over tabs vs whitespace or the various bracing style. These previous arguments were mainly eliminated with tools like `clang-format` that allowed those rules to become part of your style guide. Anyone who has been using clang-format in a large team over the last couple of years knows that we don't have those religious arguments any more, and code reviews are more productive.

https://www.youtube.com/watch?v=fv--IKZFVO8
https://mariusbancila.ro/blog/2018/11/23/join-the-east-const-revolution/
https://www.youtube.com/watch?v=z6s6bacI424

The purpose of this revision is to try to do the same for the East/West const discussion. Move the debate into the style guide and leave it there!

In addition to the new `ConstStyle: Right` or `ConstStyle: Left` there is an additional command-line argument `--const-style=left/right` which would allow an individual developer to switch the source back and forth to their own style for editing, and back to the committed style before commit. (you could imagine an IDE might offer such a switch)

The revision works by implementing a separate pass of the Annotated lines much like the SortIncludes and then create replacements for constant type declarations.

Differential Revision: https://reviews.llvm.org/D69764

[gn build] Port a44ab1702539

[InlineAdvisor] Use one single quote

[ORC] Minor renaming and typo fixes (NFC)

Two typos, one unsused include and some leftovers from the TargetProcessControl -> ExecutorProcessControl renaming

Reviewed By: xgupta

Differential Revision: https://reviews.llvm.org/D110260

[llvm] Replace tab with spaces in one test

Also use just one space after comma. Now grepping for ', offset '
in llvm/test actually finds the test for `offset`.

No behavior change.

[llvm] Fix a copy-pasto

We should use IMAGE_REL_I386_SECREL in the i386 section of this file.

IMAGE_REL_I386_SECREL and IMAGE_REL_AMD64_SECREL have the same
numeric value 0xB, so this doesn't change behavior.

[ORC][examples] Export exectuable symbols explicitly in LLJITWithExecutorProcessControl

Functions in static code that should be callable from JITed code must be exported. For dynamic libraries extern functions are exported by default. For exectuables, linkers usually strip them away unless we explicitly ask for keeping them.

Reviewed By: xgupta

Differential Revision: https://reviews.llvm.org/D110345

[mlir][sparse] Moved a conditional from the RT library to the generated MLIR.

When generating code to add an element to SparseTensorCOO (e.g., when doing dense=>sparse conversion), we used to check for nonzero values on the runtime side, whereas now we generate MLIR code to do that check.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D110121

[lldb] Show fix-it applied even if expression didn't evaluate succesfully

If we applied a fix-it before evaluating an expression and that
expression didn't evaluate correctly, we should still tell users about
the fix-it we applied since that may be the reason why it didn't work
correctly.

Differential Revision: https://reviews.llvm.org/D109908

[flang] Represent (parentheses around derived types)

The strongly typed expression representation classes supported
a representation of parentheses only around intrinsic types
with specific kinds.  Parentheses around derived type variables
must also be preserved so that expressions may be distinguished
from variables; this distinction matters for actual arguments &
construct associations.

Differential Revision: https://reviews.llvm.org/D110355

[Polly] Remove -polly-opt-fusion option.

The name of the option is misleading and has been renamed by isl to
"serialize-sccs". Instead of also renaming the option, remove it.
The option is still accessible using

    -polly-isl-arg=--no-schedule-serialize-sccs

Support: Add closing namespace comment in raw_ostream_test.cpp, NFC

The closing namespace comment prevents clang-format from dropping a
blank line after the final test. Also add in a blank line (which
simplifies merging/rebasing/etc. WIP patches).

Revert "[Polly] Implement user-directed loop distribution/fission."

This reverts commit 52c30adc7dfe6334b71adf256d81f70e7b976143 which
breaks the build when NDEBUG is defined.

[clang-format] NFC ensure Penality variable is passed back with Fixes

Fix set but not used warning

[clangd] Support `#pragma mark` in the outline

Xcode uses `#pragma mark -` to draw a divider in the outline view
and `#pragma mark Note` to add `Note` in the outline view. For more
information, see https://nshipster.com/pragma/.

Since the LSP spec doesn't contain dividers for the symbol outline,
instead we treat `#pragma mark -` as a group with children - the
decls that come after it, implicitly terminating when the symbol's
parent ends.

The following code:

```
@implementation MyClass

- (id)init {}

- (int)foo;
@end
```

Would give an outline like

```
MyClass
        > Overrides
                    > init
        > Public Accessors
                    > foo
```

Differential Revision: https://reviews.llvm.org/D105904

[CMake] Pass llvm-readelf to CMake external builds

This matches other LLVM binary tools.

Differential Revision: https://reviews.llvm.org/D110313

[RISCV] Add more tests for (and (shl x, C2), C1) that can be improved by using a pair of shifts. NFC

These tests have C1 as a shifted mask having no leading zeros and
C3 trailing zeros. If C3 is more than C2, we can select this as
(slli (srli x, C3-C2), C3).

[RISCV] Add another isel optimization for (and (shl x, c2), c1)

Turn (and (shl x, c2), c1) -> (slli (srli x, c3-c2), c3) if c1 is a
shifted mask with no leading zeros and c3 trailing zeros where c3
is greater than c2.

RegAllocGreedy: Remove an unhelpful auto, and don't use a reference

[OpenMP] Fix data-race in new device RTL

This patch fixes a data-race observed when using the new device runtime
library. The Internal control variable for the parallel level is read in
the `__kmpc_parallel_51` function while it could potentially be written
by other threads. This causes data corruption and will cause
nondetermistic behaviour in the runtime. This patch fixes this by adding
an explicit synchronization before the region starts.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D110366

[lldb] Handle malformed qfThreadInfo reply

If the remote gdbserver's qfThreadInfo reply has a trailing comma,
GDBRemoteCommunicationClient::GetCurrentProcessAndThreadIDs will return
an empty vector of thread ids. This will cause lldb to recurse through
three functions trying to get the list of threads, until it blows its
stack and crashes.

A trailing comma is a malformed response, but it shouldn't cause lldb to
crash. This patch will return the tids received before the malformed
response.

Reviewed By: clayborg, labath

Differential Revision: https://reviews.llvm.org/D109937

[profile][fuchsia] Don't include extra NUL in log messages

Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D110361

[AMDGPU] Propagate defining src reg for AGPR to AGPR Copys

On targets that do not support AGPR to AGPR copying directly, try to find the
defining accvgpr_write and propagate its source vgpr register to the copies
before register allocation so the source vgpr register does not get clobbered.

The postrapseudos pass also attempt to propagate the defining accvgpr_write but
if the register to propagate is clobbered, it will give up and create new
temporary vgpr registers instead.

Reviewed By: rampitec

Differential Revision: https://reviews.llvm.org/D108830

[libc++][NFC] Refactor the std::reverse_view tests

Mostly, remove the global assumption that all ranges have size 8.
I should have called this out during the initial review.

[ORC] Fix file header.

Revert "[compiler-rt][profile] Add padding after binary IDs"

This reverts commit 6bc9c8dfe32cc4662f2ed9041af527f69dfff13b.

Reverted because this broke some PPC buildbots.

[CSSPGO] Do not pass -fpseudo-probe-for-profiling to the linker.

The correponding linker switch has been removed by https://reviews.llvm.org/D110209, so do not pass it in clang.

Reviewed By: wenlei

Differential Revision: https://reviews.llvm.org/D110371

[RISCV] Limit transformAddImmMulImm to prevent an infinite loop.

This fixes an issue reported in D108607.

[NFC] Refactor tests to improve readability.

Fix -fno-unwind-tables -fasynchronous-unwind-tables to emit unwind tables

This matches GCC.

Change the CC1 option to encode the unwind table level (1: needed by exceptions,
2: asynchronous) so that we can support two modes in the future.

[clang-format] Fix unittest failures with -Werror

Commit a44ab1702539 added a unit test that fails to build with
-Werror which causes build bot breaks on bots that include that
option in their build. This patch just adds the necessary casts to
silence the warnings.

[ORC] Shut down services in SimpleRemoteEPCServer.

This should have been included with ExecutorBootstrapService in 78b083dbb72,
but was accidentally left out. It give services a chance to release any
resources that they have acquired.

[compiler-rt][profile] Make corrupted-profile.c more robust

This test specifically checks that profiles are not mergeable if there's a
change in the CounterPtr in the profile header. The test manually changes
CounterPtr by explicitly calling memset on some offset into the profile file.
This test would fail if binary IDs were emitted because the offset calculation
does not take into account the binary ID sizes.

This patch updates the test to use types provided in profile/InstrProfData.inc
to make it more resistant to profile layout changes.

Differential Revision: https://reviews.llvm.org/D110277

[ThinLTO] Don't emit original GUID for locals to distributed indexes

In ThinLTO for locals we normally compute the GUID from the name after
prepending the source path to get a unique global id. SamplePGO indirect
call profiles contain the target GUID without this uniquification,
however (unless compiling with -funique-internal-linkage-names).
Therefore, the index contains the original GUID of the local symbols
(without module path prepended to uniquify), in order to correctly
handle the call edges added for these indirect call profile targets
with SamplePGO.

We were emitting these to the combined index when writing it out as
bitcode, which is unnecessary and causes overhead when writing out the
indexes for distributed backends. The only use of the original GUID name
is in the thin link. Suppress it in that case. This reduced the thin
link time for a large distributed build by about 7%, and the aggregate
size of the serialized indexes by over 2%.

Continue to print it when writing out the full index, since that is just
used for debugging and testing.

Update a distributed thinlto index test to contain a local and ensure
that we don't get a COMBINED_ORIGINAL_NAME record.

Differential Revision: https://reviews.llvm.org/D110296

[ThinLTO] Fix bot failures

Fix bot failures after 2c1defeee40cf643ea6f0fa5e01164c9a4c48c30. The new
GUID I added isn't matching because it is a local with the source path
prepended. There isn't much use in matching the GUID's exactly anyway,
so remove those from the patterns.

[mlir][linalg] Merge all tiling passes into a single one.

Passes such as `linalg-tile-to-tiled-loop` are merged into `linalg-tile`.

Differential Revision: https://reviews.llvm.org/D110214

[mlir][linalg] Support tile+peel with TiledLoopOp

Only scf.for was supported until now.

Differential Revision: https://reviews.llvm.org/D110220

Add missing storageType to AttrDef to ODS

This is only noticeable when using an attribute across dialects I think.
Previously the namespace would be ommited, but it wouldn't matter as
long as the generated code stays within a single namespace.

Differential Revision: https://reviews.llvm.org/D110367

[Polly] Implement user-directed loop distribution/fission.

This is a simple version without the possibility to define distribute
points or followup-transformations. However, it is the first
transformation that has to check whether the transformation is correct.

It interprets the same metadata as the LoopDistribute pass.

Re-apply after revert in c7bcd72a38bcf99e03e4651ed5204d1a1f2bf695 with
fix: Take isBand out of #ifndef NDEBUG since it now is used
unconditionally.

[Driver] Default Generic_GCC x86 to -fasynchronous-unwind-tables

to match GCC and Clang's own x86-64.

[AMDGPU] Legalize initialized LDS variables

We don't allow an initializer for LDS variables
and there is an early abort during instruction
selection. This patch legalizes them by ignoring
the init values. During assembly emission, proper
error reporting already exists for such instances.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D109901

DebugInfo: Implement the -gsimple-template-names functionality

This excludes certain names that can't be rebuilt from the available
DWARF:

* Atomic types - no DWARF differentiating int from atomic int.
* Vector types - enough DWARF (an attribute on the array type) to do
  this, but I haven't written the extra code to add the attributes
  required for this
* Lambdas - ambiguous with any other unnamed class
* Unnamed classes/enums - would need column info for the type in
  addition to file/line number
* noexcept function types - not encoded in DWARF

[ORC] Introduce EPCGenericDylibManager / SimpleExecutorDylibManager.

EPCGenericDylibManager provides an interface for loading dylibs and looking up
symbols in the executor, implemented using EPC-calls to functions in the
executor.

SimpleExecutorDylibManager is an executor-side service that provides the
functions used by EPCGenericDylibManager.

SimpleRemoteEPC is updated to use an EPCGenericDylibManager instance to
implement the ExecutorProcessControl loadDylib and lookup methods. In a future
commit these methods will be removed, and clients updated to use
EPCGenericDylibManagers directly.

[gn build] Port a2c1cf09dfaa

[TableGen] Allow targets to entirely ignore Psets for registers

Tablegen currently expects targets to have at least one
pressure set for every broader register category. AMDGPU's
VGPR or AGPR, for instance, seemed to work correctly without
any pset, though we have forced one for each type to avoid
the assertion in computeRegUnitSets. However, psets can not
be entirely empty. At least one set is mandatory for every
target. This patch bypasses the assertion for the classes
when GeneratePressureSet is zero while ensuring the
RegUnitSets are not empty.

Reviewed By: arsenm, rampitec

Differential Revision: https://reviews.llvm.org/D110305

[Sanitizers] intercept ttyent api on FreeBSD.

and ttyentpath separately on NetBSD.

Reviewed By: vitalybuka

Differential Revision: https://reviews.llvm.org/D109843

[docs] Document the --print-passes flag in opt.

Reviewed By: aeubanks, asbirlea

Differential Revision: https://reviews.llvm.org/D109663

[ORC] Rename ExecutorAddress to ExecutorAddr.

Removing the 'ess' suffix improves the ergonomics without sacrificing clarity.
Since this class is likely to be used more frequently in the future it's worth
some short term pain to fix this now.

Remove non-portable directory separator from test

DebugInfo: STN: Handle unreconstitutable types in function types

Fix that same path separator issue again...

[PowerPC] SemaChecking for darn family of builtins

The __darn family of builtins are only available on Pwr9,
and only __darn_32 is available on both 64 and 32 bit, while the rest
are only available on 64 bit. The patch adds sema checking
for these builtins and separate the __darn_32's 32 bit
test cases.

Differential revision: https://reviews.llvm.org/D110282

[ORC-RT] Rename ExecutorAddress to ExecutorAddr.

This is an ORC-runtime counterpart to LLVM commit ef391df2b63, and the
motivation is the same: to move to a shorter name to improve the ergonomics of
this type before it's more widely adopted.