platform/upstream/llvm.git
2 years ago[NFC] Rename Instrinsic to Intrinsic
David Green [Mon, 25 Apr 2022 17:13:23 +0000 (18:13 +0100)]
[NFC] Rename Instrinsic to Intrinsic

2 years ago[test] Remove legacy PM pipeline test
Arthur Eubanks [Mon, 25 Apr 2022 17:01:17 +0000 (10:01 -0700)]
[test] Remove legacy PM pipeline test

The legacy PM for the optimization pipeline is deprecated and in the process of being cleaned up.

2 years ago[demangler][NFC] OperatorInfo table unit test
Nathan Sidwell [Fri, 8 Apr 2022 13:55:31 +0000 (06:55 -0700)]
[demangler][NFC] OperatorInfo table unit test

Placing a run-once test inside the operator lookup function caused
problems with the thread sanitizer. See D122975.

Break out the operator table into a member variable, and move the test
to the unit test machinery.

Reviewed By: dblaikie

Differential Revision: https://reviews.llvm.org/D123390

2 years ago[flang] Avoid global name conflict when BIND(C,NAME=) is used
Peter Klausler [Fri, 22 Apr 2022 21:06:54 +0000 (14:06 -0700)]
[flang] Avoid global name conflict when BIND(C,NAME=) is used

At the top level of program units in a source file, two subprograms
are allowed to have the same name if at least one of them has a
distinct interoperable binding name.  F18's symbol table requires
(most) symbols in a scope to have distinct names, though.  Solve
by using compiler-created names for the symbols of global scope
subprograms that have interoperable binding names.

Differential Revision: https://reviews.llvm.org/D124295

2 years ago[flang] Disambiguate F(X)=Y case where F is a function returning a pointer
Peter Klausler [Fri, 22 Apr 2022 21:16:16 +0000 (14:16 -0700)]
[flang] Disambiguate F(X)=Y case where F is a function returning a pointer

F(X)=Y may be initially parsed as a statement function definition; an
existing pass will detect statement functions that should be rewritten
into assignment statemets with array element references as their
left-hand side variables.  However, F() may also be a reference to a
function that returns a data pointer, and f18 did not handle this
case correctly.

The right fix is to rewrite the parse tree for F(X)=Y into an assignment
to a function reference result.  The cases that are actually assignments
to array elements -- including all of the cases previously handled --
will have their left-hand sides converted to array element references
later by another existing rewriting step.

Differential Revision: https://reviews.llvm.org/D124299

2 years ago[RISCV] Fix incorrect policy implement for unmasked vslidedown and vslideup.
Zakk Chen [Mon, 25 Apr 2022 15:44:55 +0000 (08:44 -0700)]
[RISCV] Fix incorrect policy implement for unmasked vslidedown and vslideup.

vslideup works by leaving elements 0<i<OFFSET undisturbed.
so it need the destination operand as input for correctness
regardless of policy. Add a operand to indicate policy.

We also add policy operand for unmaksed vslidedown to keep the interface consistent with vslideup
because vslidedown have only undisturbed at 0<i<vstart but user have no way to control of vstart.

Reviewed By: rogfer01, craig.topper

Differential Revision: https://reviews.llvm.org/D124186

2 years ago[X86] combineX86ShuffleChain - don't fold to truncate(concat(V1,V2)) if it was alread...
Simon Pilgrim [Mon, 25 Apr 2022 15:51:57 +0000 (16:51 +0100)]
[X86] combineX86ShuffleChain - don't fold to truncate(concat(V1,V2)) if it was already a PACK op

Fixes #55050

2 years ago[AMDGPU] Lit test pre-commit changes (NFC)
Christudasan Devadasan [Mon, 25 Apr 2022 15:47:33 +0000 (21:17 +0530)]
[AMDGPU] Lit test pre-commit changes (NFC)

Run line change needed for an upcoming patch.

2 years ago[AMDGPU] Regenerate lit test pattern (NFC).
Christudasan Devadasan [Mon, 25 Apr 2022 15:32:28 +0000 (21:02 +0530)]
[AMDGPU] Regenerate lit test pattern (NFC).

2 years agoCODE_OWNERS: update information for builtins
Saleem Abdulrasool [Fri, 22 Apr 2022 14:34:25 +0000 (14:34 +0000)]
CODE_OWNERS: update information for builtins

Mark myself as the code owner for the builtins library as per the
discussion at
https://discourse.llvm.org/t/code-owner-for-compiler-rt-builtins-library.

Differential Revision: https://reviews.llvm.org/D124263
Reviewed By: aaron.ballman, lattner

2 years agoClear temporary file in test, buildbot appears to be reusing an old one.
Sam McCall [Mon, 25 Apr 2022 15:25:38 +0000 (17:25 +0200)]
Clear temporary file in test, buildbot appears to be reusing an old one.

https://lab.llvm.org/buildbot/#/builders/214/builds/903/steps/6/logs/FAIL__Clang__pch-with-module_m

2 years agoRevert "[AMDGPU] Use d16 flag for image.sample instructions"
Piotr Sobczak [Mon, 25 Apr 2022 15:18:49 +0000 (17:18 +0200)]
Revert "[AMDGPU] Use d16 flag for image.sample instructions"

This reverts commit d1762fc454c0d7ee0bcffe87e798f67b6c43c1d2.

Reverting D124232 as the buildbot reported some errors in sanitizers.

2 years ago[clang][dataflow] Fix `Environment::join`'s handling of flow-condition merging.
Yitzhak Mandelbaum [Thu, 14 Apr 2022 13:42:02 +0000 (13:42 +0000)]
[clang][dataflow] Fix `Environment::join`'s handling of flow-condition merging.

The current implementation mutates the environment as it performs the
join. However, that interferes with the call to the model's `merge` operation,
which can modify `MergedEnv`. Since any modifications are assumed to apply to
the post-join environment, providing the same environment for both is
incorrect. This mismatch is a particular concern for joining the flow
conditions, where modifications in the old environment may not be propagated to
the new one.

Differential Revision: https://reviews.llvm.org/D124104

2 years agoReapply D124184, [DebugInfo][InstrRef] Add a size operand to DBG_PHI
Jeremy Morse [Mon, 25 Apr 2022 14:23:57 +0000 (15:23 +0100)]
Reapply D124184, [DebugInfo][InstrRef] Add a size operand to DBG_PHI

This was applied in fda4305e53784, reverted in 13815e8cbf8d49, the problem
was that fp80 X86 registers that were spilt to the stack aren't expected by
LiveDebugValues. It pre-allocates a position number for all register sizes
that can be spilt, and 80 bits isn't exactly common.

The solution is to scan the register classes to find any unrecognised
register sizes, adn pre-allocate those position numbers, avoiding a later
assertion.

2 years ago[Libomptarget] Use entry name for global info
Joseph Huber [Mon, 25 Apr 2022 12:28:30 +0000 (08:28 -0400)]
[Libomptarget] Use entry name for global info

Currently, globals on the device will have an infinite reference count
and an unknown name when using `LIBOMPTARGET_INFO` to print the mapping
table. We already store the name of the global in the offloading entry
so we should be able to use it, although there will be no source
location. To do this we need to create a valid `ident_t` string from a
name only.

Reviewed By: tianshilei1992

Differential Revision: https://reviews.llvm.org/D124381

2 years ago[SimpleLoopUnswitch] Enable freezing of conditions by default.
Florian Hahn [Mon, 25 Apr 2022 13:26:41 +0000 (14:26 +0100)]
[SimpleLoopUnswitch] Enable freezing of conditions by default.

This fixes a series of mis-compiles by SimpleLoopUnswitch.

My measurements showed no performance regression with -O3 on AArch64
in SPEC2006, SPEC2017 and a set of internal benchmarks.

Fixes #50387, #50430

Depends on D124251.

Reviewed By: nikic, aqjune

Differential Revision: https://reviews.llvm.org/D124252

2 years ago[libc++] Fix C++03 with the unstable ABI enabled
Nikolas Klauser [Mon, 25 Apr 2022 13:17:12 +0000 (15:17 +0200)]
[libc++] Fix C++03 with the unstable ABI enabled

2 years agoRevert "[DebugInfo][InstrRef] Add a size operand to DBG_PHI"
Jeremy Morse [Mon, 25 Apr 2022 13:06:12 +0000 (14:06 +0100)]
Revert "[DebugInfo][InstrRef] Add a size operand to DBG_PHI"

This reverts commit fda4305e5378478051be225248bfe9c1d401d938.

Green dragon has spotted a problem -- it's understood, but might be fiddly
to fix, reverting in the meantime.

2 years agoRevert "[sanitizer] Use canonical syscalls everywhere"
Nico Weber [Mon, 25 Apr 2022 12:48:57 +0000 (08:48 -0400)]
Revert "[sanitizer] Use canonical syscalls everywhere"

This reverts commit 34b676eb60ca1fa012068d161633f268d8ea7e6c.
Speculative, might have caused test problems on Android.

2 years ago[DebugInfo][InstrRef] Add a size operand to DBG_PHI
Jeremy Morse [Thu, 21 Apr 2022 13:39:39 +0000 (14:39 +0100)]
[DebugInfo][InstrRef] Add a size operand to DBG_PHI

DBG_PHI instructions can refer to stack slots, to indicate that multiple
values merge together on control flow joins in that slot. This is fine --
however the slot might be merged at a later date with a slot of a different
size. In doing so, we lose information about the size the eliminated PHI.
Later analysis passes have to guess.

Improve this by attaching an optional "bit size" operand to DBG_PHI, which
only gets added for stack slots, to let us know how large a size the value
on the stack is.

Differential Revision: https://reviews.llvm.org/D124184

2 years ago[Frontend] shrink in-memory PCH buffers to fit
Sam McCall [Fri, 22 Apr 2022 09:59:38 +0000 (11:59 +0200)]
[Frontend] shrink in-memory PCH buffers to fit

After building a PCH, the vector capacity is on average ~1/3 unused.
If we're going to keep it in memory for a while, reallocate to the right size.
Take care to do this once clang is destroyed so that we can reuse its
memory rather than requesting more.

Differential Revision: https://reviews.llvm.org/D124242

2 years ago[InstCombine] Remove redundant unsigned underflow fold (NFCI)
Nikita Popov [Mon, 25 Apr 2022 11:18:31 +0000 (13:18 +0200)]
[InstCombine] Remove redundant unsigned underflow fold (NFCI)

This is now handled as a combination of two other folds:
(A+B) <= A & (A+B) != 0  -->  (A+B)-1 < A
(A+B)-1 < A  -->  -B < A

2 years ago[AMDGPU] Use d16 flag for image.sample instructions
Mariusz Sikora [Mon, 25 Apr 2022 11:57:27 +0000 (12:57 +0100)]
[AMDGPU] Use d16 flag for image.sample instructions

Image.sample instruction can be forced to return half type instead of
float when d16 flag is enabled.

This patch adds new pattern in InstCombine to detect if output of
image.sample is used later only by fptrunc which converts the type
from float to half. If pattern is detected then fptrunc and image.sample
are combined to single image.sample which is returning half type.
Later in Lowering part d16 flag is added to image sample intrinsic.

Differential Revision: https://reviews.llvm.org/D124232

2 years ago[X86][AMX] Add test case for D124378.
Luo, Yuanke [Mon, 25 Apr 2022 11:47:21 +0000 (19:47 +0800)]
[X86][AMX] Add test case for D124378.

2 years ago[flang][driver] Add support for generating executables
Andrzej Warzynski [Thu, 17 Mar 2022 16:24:54 +0000 (16:24 +0000)]
[flang][driver] Add support for generating executables

This patch adds 2 missing items required for `flang-new` to be able to
generate executables:

1. The Fortran_main runtime library, which implements the main entry
   point into Fortran's `PROGRAM` in Flang,

2. Extra linker flags to include Fortran runtime libraries (e.g.
   Fortran_main).

Fortran_main is the bridge between object files generated by Flang and
the C runtime that takes care of program set-up at system-level. For
every Fortran `PROGRAM`, Flang generates the `_QQmain` function.
Fortran_main implements the C `main` function that simply calls
`_QQmain`.

Additionally, "<driver-path>/../lib" directory is added to the list of
search directories for libraries. This is where the required runtime
libraries are currently located. Note that this the case for the build
directory. We haven't considered installation directories/targets yet.

With this change, you can generate an executable that will print `hello,
world!` as follows:

```bash
$ cat hello.f95
PROGRAM HELLO
  write(*, *) "hello, world!"
END PROGRAM HELLO
$ flang-new -flang-experimental-exec hello.f95
./a.out
hello, world!
```

NOTE 1: Fortran_main has to be a static library at all times. It invokes
`_QQmain`, which is the main entry point generated by Flang for the
given input file (you can check this with `flang-new -S hello.f95 -o - |
grep "Qmain"`). This means that Fortran_main has an unresolved
dependency at build time. The linker will allow this for a static
library. However, if Fortran_main was a shared object, then the linker
will produce an error: `undefined symbol: `_QQmain`.

NOTE 2: When Fortran runtime libraries are generated as shared libraries
(excluding Fortran_main, which is always static), you will need to
tell the dynamic linker (by e.g. tweaking LD_LIBRARY_PATH) where to look
for them when invoking the executables. For example:
```bash
LD_LIBRARY_PATH=$LD_LIBRARY_PATH:<flang-build-dir>/lib/ ./a.out
```

NOTE 3: This feature is considered experimental and currently guarded
with a flag: `-flang-experimental-exec`.

Differential Revision: https://reviews.llvm.org/D122008

[1] https://github.com/flang-compiler/f18-llvm-project

CREDITS: Fortran_main was originally written by Eric Schweitz, Jean
Perier, Peter Klausler and Steve Scalpone in the fir-dev` branch in [1].

Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Co-authored-by: Peter Klausler <pklausler@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Steve Scalpone <sscalpone@nvidia.com
2 years ago[SimpleLoopUnswitch] Run LICM for nested unswitching tests.
Florian Hahn [Mon, 25 Apr 2022 11:49:07 +0000 (12:49 +0100)]
[SimpleLoopUnswitch] Run LICM for nested unswitching tests.

When enabling freeze-loop-unswitch-cond the inserted freeze instruction
may block unswitching of parent loops if they get inserted in a block in
the parent loop (as the llvm::Loop-based invariance check only checks
 whether an instruction is in a loop block or not).

In the optimization pipeline, LICM is responsible to hoist out loop
invariant instructions to enable further unswitching. Also run LICM for
nested unswitching tests in preparation for flipping the default of
freeze-loop-unswitch-cond.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D124251

2 years ago[X86][AVX] Add shuffle test case for Issue #55066
Simon Pilgrim [Mon, 25 Apr 2022 11:17:23 +0000 (12:17 +0100)]
[X86][AVX] Add shuffle test case for Issue #55066

2 years ago[InstCombine] Fold (X != 0) & (Y u>= X)
Nikita Popov [Fri, 22 Apr 2022 13:14:49 +0000 (15:14 +0200)]
[InstCombine] Fold (X != 0) & (Y u>= X)

This adds the De Morgan conjugated fold for the existing
(X == 0) | (Y u< X) fold.

Proof: https://alive2.llvm.org/ce/z/3Me3JQ

2 years ago[lld/mac] Revert libcompiler_rt.dylib version check change
Nico Weber [Mon, 25 Apr 2022 01:10:55 +0000 (21:10 -0400)]
[lld/mac] Revert libcompiler_rt.dylib version check change

This reverts D117925 since it's no longer needed after D124336.

Differential Revision: https://reviews.llvm.org/D124354

2 years ago[InstCombine] Fold X + Y + C u< X
Nikita Popov [Mon, 25 Apr 2022 10:08:30 +0000 (12:08 +0200)]
[InstCombine] Fold X + Y + C u< X

This is a variation on the X + Y u< X fold with an extra constant.
Proof: https://alive2.llvm.org/ce/z/VNb8pY

2 years ago[InstCombine] Add additional tests for X + Y + C u< X (NFC)
Nikita Popov [Mon, 25 Apr 2022 10:47:03 +0000 (12:47 +0200)]
[InstCombine] Add additional tests for X + Y + C u< X (NFC)

We don't actually need the limitation for C == -1, so update test
naming accordingly, and also test a non-uniform vector constant.

2 years ago[LoopInstSimplify] Ignore users in unreachable blocks. PR55072
Max Kazantsev [Mon, 25 Apr 2022 10:33:57 +0000 (17:33 +0700)]
[LoopInstSimplify] Ignore users in unreachable blocks. PR55072

Logic in this pass assumes that all users of loop instructions are
either in the same loop or are LCSSA Phis. In fact, there can also
be users in unreachable blocks that currently break assertions.
Such users don't need to go to the next round of simplifications.

Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D124368

2 years ago[LangRef] Limit readnone,read/writeonly to memory visible outside the fn
Florian Hahn [Mon, 25 Apr 2022 10:32:50 +0000 (11:32 +0100)]
[LangRef] Limit readnone,read/writeonly to memory visible outside the fn

This patch unifies the wording used for readnone, readonly and writeonly
attributes. The definitions now more specifically refer to memory visible
outside the function

The motivation for the clarification is D123473.

Reviewed By: nlopes

Differential Revision: https://reviews.llvm.org/D124124

2 years ago[gn build] Port 042dc3c46d73
LLVM GN Syncbot [Mon, 25 Apr 2022 10:26:57 +0000 (10:26 +0000)]
[gn build] Port 042dc3c46d73

2 years ago[mlir][LLVM] Support opaque pointers in `llvm.mlir.addressof`
Markus Böck [Mon, 25 Apr 2022 10:23:09 +0000 (12:23 +0200)]
[mlir][LLVM] Support opaque pointers in `llvm.mlir.addressof`

The verifier of llvm.mlir.addressof did not properly account for opaque pointers, that is, the pointer type not having an element type equal to the type of the referenced global or function. This patch fixes that by skipping the test for the element type if the pointer is opaque.

Differential Revision: https://reviews.llvm.org/D124333

2 years ago[libc++] add zip_view and views::zip for C++23
Hui Xie [Mon, 25 Apr 2022 10:18:54 +0000 (12:18 +0200)]
[libc++] add zip_view and views::zip for C++23

- add zip_view and views::zip for C++23
- added unit tests
- implemented section 5.6 (zip) in P2321R2

I used clang-format to format the files but they look nothing like the rest of the code base. Manually indenting each line to match the styles sounds like an impossible task. Is there any clang-format file which can format it reasonable similar to the rest of the code base so that I can manually format the rest lines that look weird?

Reviewed By: ldionne, #libc, philnik, var-const

Spies: Mordante, philnik, libcxx-commits, mgorny

Differential Revision: https://reviews.llvm.org/D122806

2 years ago[InstCombine] Add tests for X + Y - 1 u< X (NFC)
Nikita Popov [Mon, 25 Apr 2022 10:20:56 +0000 (12:20 +0200)]
[InstCombine] Add tests for X + Y - 1 u< X (NFC)

2 years ago[Test] Simplify test for PR55072
Max Kazantsev [Mon, 25 Apr 2022 10:20:08 +0000 (17:20 +0700)]
[Test] Simplify test for PR55072

2 years ago[Serialization] write expr dependence bits as a single integer
Sam McCall [Fri, 22 Apr 2022 12:00:36 +0000 (14:00 +0200)]
[Serialization] write expr dependence bits as a single integer

When exprs are written unabbreviated:
  - these were encoded as 5 x vbr6 = 30 bits
  - now they fit exactly into a one-chunk vbr = 6 bits

clangd --check=clangd/AST.cpp reports ~1% reduction in PCH size
(42826720->42474460)

Differential Revision: https://reviews.llvm.org/D124250

2 years ago[mlir][vector] insert `alloca`s outside of loops
Alex Zinenko [Mon, 25 Apr 2022 08:25:59 +0000 (10:25 +0200)]
[mlir][vector] insert `alloca`s outside of loops

After https://reviews.llvm.org/D119743 added the `AutomaticAllocationScope`
trait to loop-like constructs, the vector transfer full/partial splitting pass
started inserting allocations for temporaries within the closest loop rather
than the closest function (or other allocation scope such as `async.execute`).
While this is correct as long as the lowered code takes care of automatic
deallocation at the end of each iteration of the loop, this interferes with
downstream optimizations that expect `alloca`s to be at the function level.
Step over loops when looking for the closest allocation scope in vector
transfer full/partial splitting pass thus restoring the original behavior.

Reviewed By: hanchung

Differential Revision: https://reviews.llvm.org/D124366

2 years ago[MemorySSA] Use -simple-loop-unswitch instead of -loop-unswitch in test.
Florian Hahn [Mon, 25 Apr 2022 08:22:45 +0000 (09:22 +0100)]
[MemorySSA] Use -simple-loop-unswitch instead of -loop-unswitch in test.

2 years ago[runtimes] [CMake] Rename a cmake variable missed in b3df14b6c98702ce50401fd039852787...
Martin Storsjö [Mon, 25 Apr 2022 06:03:00 +0000 (09:03 +0300)]
[runtimes] [CMake] Rename a cmake variable missed in b3df14b6c98702ce50401fd039852787373e4676

This was missed as this check only is executed if linking doesn't
work without it.

2 years ago[Test] Add test for PR55072
Max Kazantsev [Mon, 25 Apr 2022 08:18:51 +0000 (15:18 +0700)]
[Test] Add test for PR55072

2 years ago[MemorySSA] Generate check lines for test.
Florian Hahn [Mon, 25 Apr 2022 08:02:42 +0000 (09:02 +0100)]
[MemorySSA] Generate check lines for test.

This is to ensure we produce the same code when switching to
SimpleLoopUnswitch.

2 years ago[mlir][Bazel] Add missing dependencies.
Adrian Kuegel [Mon, 25 Apr 2022 07:16:16 +0000 (09:16 +0200)]
[mlir][Bazel] Add missing dependencies.

When building with layering_check enabled, there needs to be a
dependency for each header include.

2 years ago[flang] fix LBOUND lowering with KIND and no DIM arguments
Jean Perier [Mon, 25 Apr 2022 07:20:07 +0000 (09:20 +0200)]
[flang] fix LBOUND lowering with KIND and no DIM arguments

The lowering code was mistakenly assuming that the second argument
in the signature provided by semantics is the DIM argument. This
caused calls with a KIND argument but no DIM to be lowered as if the
KIND argument was DIM.

Differential Revision: https://reviews.llvm.org/D124243

2 years ago[mlir][LLVM] Support opaque pointers in data layout entries
Markus Böck [Mon, 25 Apr 2022 07:14:18 +0000 (09:14 +0200)]
[mlir][LLVM] Support opaque pointers in data layout entries

This is likely preferable to having it crash if one were to specify an opaque pointer type, and the actual element type is unused either way.

Differential Revision: https://reviews.llvm.org/D124334

2 years agoFix an issue in comment. NFC
Jun Zhang [Mon, 25 Apr 2022 04:43:49 +0000 (12:43 +0800)]
Fix an issue in comment. NFC

I think the author renamed the function but forgot to update the
comment.
Signed-off-by: Jun Zhang <jun@junz.org>
2 years ago[OpenMP][IRBuilder] Handle unexcuted EXPECT_FALSE
Shraiysh Vaishay [Mon, 25 Apr 2022 02:53:58 +0000 (08:23 +0530)]
[OpenMP][IRBuilder] Handle unexcuted EXPECT_FALSE

This patch addresses the comment about unexecuted test in D122371.

Reviewed By: probinson

Differential Revision: https://reviews.llvm.org/D123920

2 years ago[InstCombine] Complete folding of fneg-of-fabs
Chenbing Zheng [Mon, 25 Apr 2022 01:48:43 +0000 (09:48 +0800)]
[InstCombine] Complete folding of fneg-of-fabs

This patch add a function foldSelectWithFCmpToFabs, and do more combine for
fneg-of-fabs.
With 'nsz':
fold (X <  +/-0.0) ? X : -X or (X <= +/-0.0) ? X : -X to -fabs(x)
fold (X >  +/-0.0) ? X : -X or (X >= +/-0.0) ? X : -X to -fabs(x)

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D123830

2 years ago[libcxx] random_device, use arc4random() on FreeBSD, NetBSD and DragonFlyBSD
Brad Smith [Mon, 25 Apr 2022 01:45:49 +0000 (21:45 -0400)]
[libcxx] random_device, use arc4random() on FreeBSD, NetBSD and DragonFlyBSD

Reviewed By: ldionne, emaste, dim

Differential Revision: https://reviews.llvm.org/D122628

2 years ago[clang-format] Refactor determineStarAmpUsage NFC
sstwcw [Sun, 24 Apr 2022 22:30:22 +0000 (22:30 +0000)]
[clang-format] Refactor determineStarAmpUsage NFC

There was some duplicate code in determineStarAmpUsage and
determinePlusMinusCaretUsage

Now a `-` or `+` following `;`, `sizeof`, `co_await`, or `delete` is
regarded as a unary operator.

Now a `*` or `&` following `case` is also a unary operator.

Reviewed By: curdeius, MyDeveloperDay, HazardyKnusperkeks

Differential Revision: https://reviews.llvm.org/D121754

2 years agoMove test/.gitattributes to clang-tools-extra/test
Aaron Puchert [Sun, 24 Apr 2022 20:30:25 +0000 (22:30 +0200)]
Move test/.gitattributes to clang-tools-extra/test

It was probably accidentally added there, see the discussion on change
D97625. It is certainly without effect, to quote gitattributes(5):

    When deciding what attributes are assigned to a path, Git consults
    [...], `.gitattributes` file in the same directory as the path in
    question, and its parent directories up to the toplevel of the work
    tree [...]

Running `git check-attr -a` on the files in question shows that now the
settings are indeed effective whereas before they were not.

Lastly, lit ignores the file like any dotfile, see getTestsInDirectory
of FileBasedTest in llvm/utils/lit/lit/formats/base.py. This can be
verified with `llvm-lit --show-tests clang-tools-extra/test`.

2 years ago[flang] Do not create arith.extui with same from/to type
Valentin Clement (バレンタイン クレメン) [Sun, 24 Apr 2022 18:37:27 +0000 (20:37 +0200)]
[flang] Do not create arith.extui with same from/to type

In some case the lowering of `ichar` is generating an `arith.extui` operation
with the same from/to type. This operation do not accept from/to types to be
the same. If the from/to types are identical, we do not generate the extra
operation.

Reviewed By: jeanPerier

Differential Revision: https://reviews.llvm.org/D124107

2 years agorun-clang-tidy: Fix infinite loop on windows
Jonas Toth [Sun, 24 Apr 2022 15:16:36 +0000 (17:16 +0200)]
run-clang-tidy: Fix infinite loop on windows

`find_compilation_database` checked only for "/" as exit point, but on Windows, this root is impossible.
Fixes #53642

Authored By: Febbe
Reviewed By: JonasToth
Differential Revision: https://reviews.llvm.org/D119481

2 years agollvm-reduce: Don't delete triple/datalayout
Matt Arsenault [Sun, 17 Apr 2022 14:15:16 +0000 (10:15 -0400)]
llvm-reduce: Don't delete triple/datalayout

Removing these is extremely unhelpful and just adds extra hassle. This
is really finding out whether your test script uses -mtriple or
not. You can't meaningfully delete these fields, and the resulting
module defaults to the host.

2 years ago[runtimes] [CMake] Unify variable names
Petr Hosek [Thu, 10 Mar 2022 09:47:09 +0000 (11:47 +0200)]
[runtimes] [CMake] Unify variable names

Avoid repeating CMake checks across runtimes by unifying names of
variables used for results to leverage CMake caching.

Differential Revision: https://reviews.llvm.org/D110005

2 years ago[InstCombine] Add test for fold fcmp-of-copysign. nfc
Chenbing Zheng [Sun, 24 Apr 2022 09:26:15 +0000 (17:26 +0800)]
[InstCombine] Add test for fold fcmp-of-copysign. nfc

Add more tests with differert predicates.

2 years ago[analyzer] Add option for AddrSpace in core.NullDereference check
Vince Bridgers [Fri, 8 Apr 2022 11:17:14 +0000 (06:17 -0500)]
[analyzer] Add option for AddrSpace in core.NullDereference check

    This change adds an option to detect all null dereferences for
    non-default address spaces, except for address spaces 256, 257 and 258.
    Those address spaces are special since null dereferences are not errors.

    All address spaces can be considered (except for 256, 257, and 258) by
    using -analyzer-config
    core.NullDereference:DetectAllNullDereferences=true. This option is
    false by default, retaining the original behavior.

    A LIT test was enhanced to cover this case, and the rst documentation
    was updated to describe this behavior.

Reviewed By: steakhal

Differential Revision: https://reviews.llvm.org/D122841

2 years ago[RISCV] Add sched to pseudo function call instructions
wangpc [Sun, 24 Apr 2022 06:55:54 +0000 (14:55 +0800)]
[RISCV] Add sched to pseudo function call instructions

To fix llvm-mca's error of 'found an unsupported instruction
in the input assembly sequence.' caused by the lack of
scheduling info.

Pseudo function call instructions will be expanded to `auipc`
and `jalr`, so their scheduling info are the combination of
two.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D123578

2 years ago[NFC][Clang][Pragma] Remove unused variables
Senran Zhang [Sun, 24 Apr 2022 02:49:58 +0000 (10:49 +0800)]
[NFC][Clang][Pragma] Remove unused variables

Reviewed By: beanz

Differential Revision: https://reviews.llvm.org/D124339

2 years ago[OCaml][DebugInfo] Add bindings for parameter and auto variable creation
Vaivaswatha Nagaraj [Sun, 24 Apr 2022 04:07:04 +0000 (09:37 +0530)]
[OCaml][DebugInfo] Add bindings for parameter and auto variable creation

This patch extends https://reviews.llvm.org/D90831 with a few more
functions added to the OCaml debuginfo bindings.

Differential Revision: https://reviews.llvm.org/D123914

2 years ago[lld/mac] For catalyst outputs, tolerate implicitly linking against mac-only tbd...
Nico Weber [Fri, 22 Apr 2022 15:55:50 +0000 (11:55 -0400)]
[lld/mac] For catalyst outputs, tolerate implicitly linking against mac-only tbd files

Before this,

  clang empty.cc -target x86_64-apple-ios13.1-macabi \
      -framework CoreServices -fuse-ld=lld

would error out with

    ld64.lld: error: path/to/MacOSX.sdk/System/Library/Frameworks/
         CoreServices.framework/Versions/A/Frameworks/CarbonCore.framework/
         Versions/A/CarbonCore.tbd(
             /System/Library/Frameworks/
             CoreServices.framework/Versions/A/Frameworks/CarbonCore.framework/
             Versions/A/CarbonCore) is incompatible with x86_64 (macCatalyst)

Now it works, like with ld64.

Differential Revision: https://reviews.llvm.org/D124336

2 years ago[InlineCost] Set LastCallToStaticBonus in ML inlining models.
Jun Ma [Fri, 22 Apr 2022 07:08:01 +0000 (15:08 +0800)]
[InlineCost] Set LastCallToStaticBonus in ML inlining models.

This patch set LastCallToStaticBonus based on check, it has
no noticeable size reduction on an internal workload and linux kernel
with Os/Oz.

Differential Revision: https://reviews.llvm.org/D124233

2 years agoRevert "[mlir][sparse] Expose SpareTensor passes as enums instead of opaque numbers...
Nick Kreeger [Sun, 24 Apr 2022 01:14:48 +0000 (20:14 -0500)]
Revert "[mlir][sparse] Expose SpareTensor passes as enums instead of opaque numbers for vectorization and parallelization options."

This reverts commit d59cf901cbae7991f7847eb038d825efff1221ad.

Build fails on NVIDIA Sparse tests:
https://lab.llvm.org/buildbot/#/builders/61/builds/25447

2 years agoReverting accidental git-revert commits.
Nick Kreeger [Sun, 24 Apr 2022 01:12:15 +0000 (20:12 -0500)]
Reverting accidental git-revert commits.

Revert "Revert "Revert "Revert "[Concepts] Fix overload resolution bug with constrained candidates""""

This reverts commit cfc2c5905ec11a501cdfc9502f503ab494c200b6.

2 years agoReverting accidental git-revert commits.
Nick Kreeger [Sun, 24 Apr 2022 01:11:58 +0000 (20:11 -0500)]
Reverting accidental git-revert commits.

Revert "Revert "Revert "[Concepts] Fix overload resolution bug with constrained candidates"""

This reverts commit a0636b5855f5ba1f7033670dc32c956d63baaa51.

2 years agoReverting accidental git-revert commits.
Nick Kreeger [Sun, 24 Apr 2022 01:11:27 +0000 (20:11 -0500)]
Reverting accidental git-revert commits.

Revert "Revert "[Concepts] Fix overload resolution bug with constrained candidates""

This reverts commit f6af446b6625657b1b9046273f5b33bc1173a97c.

2 years agoRevert "Revert "Revert "[Concepts] Fix overload resolution bug with constrained candi...
Nick Kreeger [Sun, 24 Apr 2022 00:36:51 +0000 (19:36 -0500)]
Revert "Revert "Revert "[Concepts] Fix overload resolution bug with constrained candidates"""

This reverts commit a0636b5855f5ba1f7033670dc32c956d63baaa51.

2 years agoRevert "Revert "[Concepts] Fix overload resolution bug with constrained candidates""
Nick Kreeger [Sun, 24 Apr 2022 00:35:50 +0000 (19:35 -0500)]
Revert "Revert "[Concepts] Fix overload resolution bug with constrained candidates""

This reverts commit f6af446b6625657b1b9046273f5b33bc1173a97c.

2 years agoRevert "[Concepts] Fix overload resolution bug with constrained candidates"
Nick Kreeger [Sun, 24 Apr 2022 00:35:41 +0000 (19:35 -0500)]
Revert "[Concepts] Fix overload resolution bug with constrained candidates"

This reverts commit 807e418413a0958ad1ea862093fb262673b2afa1.

2 years ago[mlir][sparse] Expose SpareTensor passes as enums instead of opaque numbers for vecto...
Nick Kreeger [Sun, 24 Apr 2022 00:16:57 +0000 (19:16 -0500)]
[mlir][sparse] Expose SpareTensor passes as enums instead of opaque numbers for vectorization and parallelization options.

The SparseTensor passes currently use opaque numbers for the CLI, despite using an enum internally. This patch exposes the enums instead of numbered items that are matched back to the enum.

Fixes GitHub issue #53389

Reviewed by: aartbik, mehdi_amini

Differential Revision: https://reviews.llvm.org/D123876

2 years ago[Concepts] Fix overload resolution bug with constrained candidates
Roy Jacobson [Fri, 15 Apr 2022 15:58:11 +0000 (11:58 -0400)]
[Concepts] Fix overload resolution bug with constrained candidates

When doing overload resolution, we have to check that candidates' parameter types are equal before trying to find a better candidate through checking which candidate is more constrained.
This revision adds this missing check and makes us diagnose those cases as ambiguous calls when the types are not equal.

Fixes GitHub issue https://github.com/llvm/llvm-project/issues/53640

Reviewed By: erichkeane

Differential Revision: https://reviews.llvm.org/D123182

2 years ago[hwasan][test] Change -hwasan to -passes=hwasan
Fangrui Song [Sat, 23 Apr 2022 21:07:31 +0000 (14:07 -0700)]
[hwasan][test] Change -hwasan to -passes=hwasan

With the -enable-new-pm defaulting to true, the two forms have the same effect,
but the previous form uses the legacy pass manager syntax which is being phased
out.

2 years ago[Bazel] Add more mlir dependencies after D124298
Fangrui Song [Sat, 23 Apr 2022 20:06:54 +0000 (13:06 -0700)]
[Bazel] Add more mlir dependencies after D124298

The Bazel layering_check feature compiles libraries with `-fmodule-name=X
-fmodules-strict-decluse` which require #include to be in deps.

2 years ago[Bazel] Make mlir:BufferizationDialect depend on mlir:ArithmeticDialect after D124298
Fangrui Song [Sat, 23 Apr 2022 19:59:37 +0000 (12:59 -0700)]
[Bazel] Make mlir:BufferizationDialect depend on mlir:ArithmeticDialect after D124298

2 years ago[analyzer] Clean checker options from bool to DefaultBool (NFC)
Vince Bridgers [Sun, 10 Apr 2022 12:44:11 +0000 (07:44 -0500)]
[analyzer] Clean checker options from bool to DefaultBool (NFC)

A recent review emphasized the preference to use DefaultBool instead of
bool for checker options. This change is a NFC and cleans up some of the
instances where bool was used, and could be changed to DefaultBool.

Reviewed By: steakhal

Differential Revision: https://reviews.llvm.org/D123464

2 years ago[OpenMP] Use CUDA's non-RDC mode when LTO has whole program visibility
Joseph Huber [Fri, 22 Apr 2022 17:19:16 +0000 (13:19 -0400)]
[OpenMP] Use CUDA's non-RDC mode when LTO has whole program visibility

When we do LTO we consider ourselves to have whole program visibility if
every single input file we have contains LLVM bitcode. If we have whole
program visibliity then we can create a single image and utilize CUDA's
non-RDC mode by not passing `-c` to `ptxas` and ignoring the `nvlink`
job. This should be faster for some situations and also saves us the
time executing `nvlink`.

Reviewed By: tra

Differential Revision: https://reviews.llvm.org/D124292

2 years ago[libc++] Fixes concepts overload resolution.
Mark de Wever [Wed, 20 Apr 2022 16:47:13 +0000 (18:47 +0200)]
[libc++] Fixes concepts overload resolution.

D123182 fixes a bug in Clang's overload resolution. After it landed it
was discovered `basic_format_arg`'s constructors contains this bug. This
fixes the bug in libc++, unblocking D123182.

The code has been tested in combination with D123182.

Reviewed By: royjacobson, #libc

Differential Revision: https://reviews.llvm.org/D124103

2 years ago[BOLT][CMAKE] Simplify Clang/LLD identification
Amir Ayupov [Sat, 23 Apr 2022 16:16:53 +0000 (12:16 -0400)]
[BOLT][CMAKE] Simplify Clang/LLD identification

Refactor nested conditions. NFC

Reviewed By: maksfb

Differential Revision: https://reviews.llvm.org/D123861

2 years agoRevert "[clang] Adding Platform/Architecture Specific Resource Header Installation...
Ye Luo [Sat, 23 Apr 2022 14:25:54 +0000 (09:25 -0500)]
Revert "[clang] Adding Platform/Architecture Specific Resource Header Installation Targets"

Caused build failure see github issue #55002

This reverts commit 2512a875ccac158bc9b654b09e3347db167e33df.

2 years ago[C++20][Modules][Driver][HU 4/N] Add fdirectives-only mode for preprocessing output.
Iain Sandoe [Fri, 31 Jul 2020 09:44:12 +0000 (10:44 +0100)]
[C++20][Modules][Driver][HU 4/N] Add fdirectives-only mode for preprocessing output.

When the -fdirectives-only option is used together with -E, the preprocessor
output reflects evaluation of if/then/else directives.

Thus it preserves macros that are still live after such processing.
This output can be consumed by a second compilation to produce a header unit.

We automatically invoke this (with -E) when we know that the job produces a
header unit so that the preprocessed output reflects the macros that will be
defined when the binary HU is emitted.

Differential Revision: https://reviews.llvm.org/D121591

2 years ago[NFC] Sync code for upstreaming.
Eric Schweitz [Fri, 22 Apr 2022 20:57:21 +0000 (13:57 -0700)]
[NFC] Sync code for upstreaming.

Remove redundant code.

Differential Revision: https://reviews.llvm.org/D124310

2 years ago[IndVarSimplify] Replace legacy loop-unswitch with SimpleLoopUnswitch.
Florian Hahn [Sat, 23 Apr 2022 13:08:47 +0000 (14:08 +0100)]
[IndVarSimplify] Replace legacy loop-unswitch with SimpleLoopUnswitch.

Also moves the test to use the new pass manager.

2 years ago[SimpleLoopUnswitch] Use SLU version of memoryssa-threshold option.
Florian Hahn [Sat, 23 Apr 2022 13:03:32 +0000 (14:03 +0100)]
[SimpleLoopUnswitch] Use SLU version of memoryssa-threshold option.

Update the test to use the SLU version of the option.

2 years ago[SCCP] Add initial tests for freeze.
Florian Hahn [Fri, 22 Apr 2022 21:23:21 +0000 (22:23 +0100)]
[SCCP] Add initial tests for freeze.

2 years ago[SCEV] Use getConstant to construct SCEV for ConstantInt (NFC).
Florian Hahn [Sat, 23 Apr 2022 10:12:58 +0000 (11:12 +0100)]
[SCEV] Use getConstant to construct SCEV for ConstantInt (NFC).

We already know that we will construct a SCEVConstant. Directly use
getConstant, rather than going through getSCEV.

2 years ago[mlir][bufferize][NFC] Remove old references to Comprehensive Bufferize
Matthias Springer [Sat, 23 Apr 2022 08:59:47 +0000 (17:59 +0900)]
[mlir][bufferize][NFC] Remove old references to Comprehensive Bufferize

Differential Revision: https://reviews.llvm.org/D124324

2 years ago[mlir][bufferize][NFC] Clean up test cases
Matthias Springer [Sat, 23 Apr 2022 08:59:06 +0000 (17:59 +0900)]
[mlir][bufferize][NFC] Clean up test cases

Run `one-shot-bufferize` instead of `linalg-comprehensive-module-bufferize` and move some test cases to their respective dialects.

Differential Revision: https://reviews.llvm.org/D124323

2 years ago[C++20][Modules][Driver][HU 3/N] Handle foo.h with -fmodule-header and/or C++ invocation.
Iain Sandoe [Thu, 18 Feb 2021 12:23:15 +0000 (12:23 +0000)]
[C++20][Modules][Driver][HU 3/N] Handle foo.h with -fmodule-header and/or C++ invocation.

Allow an invocation like clang -fmodule-header bar.h (which will be a C++
compilation, but using a header which will be recognised as a C one).

Also  we do not want to produce:
 "treating 'c-header' input as 'c++-header' when in C++ mode"
diagnostics when the user has been specific about the intent.

Differential Revision: https://reviews.llvm.org/D121590

2 years ago[mlir][NFC] Shift a bunch of dialect includes from the .h to the .cpp
River Riddle [Fri, 22 Apr 2022 21:17:08 +0000 (14:17 -0700)]
[mlir][NFC] Shift a bunch of dialect includes from the .h to the .cpp

Now that dialect constructors are generated in the .cpp file, we can
drop all of the dependent dialect includes from the .h file.

Differential Revision: https://reviews.llvm.org/D124298

2 years ago[mlir] Generate Dialect constructors in .cpp instead of .h
River Riddle [Fri, 22 Apr 2022 21:16:06 +0000 (14:16 -0700)]
[mlir] Generate Dialect constructors in .cpp instead of .h

By generating in the .h file, we were forcing dialects to include
a lot of additional header files because:

* Fields of the dialect, e.g. std::unique_ptr<>, were unable to use
  forward declarations.
* Dependent dialects are loaded in the constructor, requiring the
  full definition of each dependent dialect (which, depending on
  the file structure of the dialect, may include the operations).

By generating in the .cpp we get much faster builds, and also
better align with the rest of the code base.

Fixes #55044

Differential Revision: https://reviews.llvm.org/D124297

2 years ago[clang][OpenMP][DebugInfo] Debug support for TLS variables present in OpenMP consruct
Alok Kumar Sharma [Thu, 14 Apr 2022 09:04:22 +0000 (14:34 +0530)]
[clang][OpenMP][DebugInfo] Debug support for TLS variables present in OpenMP consruct

In case of OpenMP programs, thread local variables can be present in
any clause pertaining to OpenMP constructs, as we know that compiler
generates artificial functions and in some cases values are passed to
those artificial functions thru parameters. For an example, if thread
local variable is present in copyin clause (testcase attached with the
patch), parameter with same name is generated as parameter to artificial
function. When user inquires the thread Local variable, its debug info
is hidden by the parameter. User never gets the actual TLS variable
when inquires it, instead gets the artificial parameter.

Current patch suppresses the debug info for such artificial parameter to
enable correct debugging of TLS variables.

Reviewed By: aprantl

Differential Revision: https://reviews.llvm.org/D123787

2 years ago[test] Don't rely on default of detect_stack_use_after_return
Vitaly Buka [Sat, 23 Apr 2022 06:28:27 +0000 (23:28 -0700)]
[test] Don't rely on default of detect_stack_use_after_return

2 years agoRevert "[test] Don't rely on default of detect_stack_use_after_return"
Vitaly Buka [Sat, 23 Apr 2022 06:25:02 +0000 (23:25 -0700)]
Revert "[test] Don't rely on default of detect_stack_use_after_return"

Revent acidentally deleted lines.

This reverts commit ff25f694731f7b9fa33f2955e751f3ccb3c35afa.

2 years ago[NFC][LAA] Match-up type sizes for possible extensions, based on actual bit-size...
Chang-Sun Lin Jr [Sat, 23 Apr 2022 06:16:20 +0000 (23:16 -0700)]
[NFC][LAA] Match-up type sizes for possible extensions, based on actual bit-size rather than rounded-up byte size.

Differential Revision: https://reviews.llvm.org/D119200

2 years ago[test] Don't rely on default of detect_stack_use_after_return
Vitaly Buka [Sat, 23 Apr 2022 02:08:56 +0000 (19:08 -0700)]
[test] Don't rely on default of detect_stack_use_after_return

2 years ago[ASan][test] Fix mismatched-new-delete in asan test unaligned_loads_and_stores.cpp
Enna1 [Sat, 23 Apr 2022 03:57:23 +0000 (11:57 +0800)]
[ASan][test] Fix mismatched-new-delete in asan test unaligned_loads_and_stores.cpp

Fix mismatched-new-delete in asan test unaligned_loads_and_stores.cpp

Reviewed By: vitalybuka, MaskRay

Differential Revision: https://reviews.llvm.org/D124268

2 years ago[Driver] Call hasFlag instead of hasArg
Akira Hatanaka [Fri, 22 Apr 2022 23:48:38 +0000 (16:48 -0700)]
[Driver] Call hasFlag instead of hasArg

`_HAS_STATIC_RTTI` should be set to 0 only by `-fno-rtti` according to the
summary of https://reviews.llvm.org/D103771.

rdar://92039243

Differential Revision: https://reviews.llvm.org/D124312

2 years ago[RISCV][test] Rename prefix and remove unused text (NFC)
Zakk Chen [Fri, 22 Apr 2022 03:01:54 +0000 (20:01 -0700)]
[RISCV][test] Rename prefix and remove unused text (NFC)

In D123418 we removed some RUN line (ex. RV32-ELEN16) but their
expected results still exist there.
Remove them and rename prefix for more descriptive.

Reviewed By: frasercrmck, asb, craig.topper

Differential Revision: https://reviews.llvm.org/D124179