platform/upstream/llvm.git
19 months ago[clang][NFC] Remove dependency on DataLayout::getPrefTypeAlignment
Guillaume Chatelet [Fri, 13 Jan 2023 13:18:45 +0000 (13:18 +0000)]
[clang][NFC] Remove dependency on DataLayout::getPrefTypeAlignment

19 months ago[SVE] Restrict SVE fixed length extload/truncstore combine to float and double types.
Paul Walker [Sun, 18 Dec 2022 16:22:39 +0000 (16:22 +0000)]
[SVE] Restrict SVE fixed length extload/truncstore combine to float and double types.

Prior to this patch we would create floating point extending load
and truncating store operations involving fp128 types, which we
cannot lower.

Fixes #58530

Differential Revision: https://reviews.llvm.org/D140318

19 months ago[X86][test] Add pre-commit test for D141657.
Freddy Ye [Fri, 13 Jan 2023 13:09:13 +0000 (21:09 +0800)]
[X86][test] Add pre-commit test for D141657.

Reviewed By: skan

Differential Revision: https://reviews.llvm.org/D141677

19 months agoRe-apply "[mlir][SparseTensor] Add a few more tests for sparse vectorization"
Quentin Colombet [Fri, 13 Jan 2023 10:32:54 +0000 (10:32 +0000)]
Re-apply "[mlir][SparseTensor] Add a few more tests for sparse vectorization"

This reverts commit 93f40c983e0adbb63cbb7c59814090134d691dd1.

Update the tests to also work on window.
The order in which the `arith.constant`s appear in the output IR is
slightly different between window and linux.
Use CHECK.*-DAG for the constants.

Original message:
These tests cover muli, xor, and, addf, subf, and addi.

The tests themselves are not that interesting, their goal is to provide
code coverage for all the types of reductions currently supported.

NFC

Differential Revision: https://reviews.llvm.org/D141369

19 months ago[RISCV] Optimize (brcond (seteq (and X, (1 << C)-1), 0))
LiaoChunyu [Thu, 12 Jan 2023 13:58:38 +0000 (21:58 +0800)]
[RISCV] Optimize (brcond (seteq (and X, (1 << C)-1), 0))

Inspired by gcc's assembly: https://godbolt.org/z/54hbzsGYn, while referring to D130203

Replace AND+IMM{32,64} with a slli.

But gcc does not handle 0xffff and 0xffffffff, which also seem to be optimizable.

The testcases copies all the bits in D130203 and adds 16, 32, and 64 bits.

Differential Revision: https://reviews.llvm.org/D141607

19 months ago[LLDB][RISCV] Add RVDC instruction support for EmulateInstructionRISCV
Emmmer [Wed, 4 Jan 2023 09:31:10 +0000 (17:31 +0800)]
[LLDB][RISCV] Add RVDC instruction support for EmulateInstructionRISCV

RVC is the RISC-V standard compressed instruction-set extension, named "C", which reduces static and dynamic code size by adding short 16-bit instruction encodings for common operations, and RVCD is the compressed "D extension".

And "D extension" is a double-precision floating-point instruction-set extension, which adds double-precision floating-point computational instructions compliant with the IEEE 754-2008 arithmetic standard.

Reviewed By: DavidSpickett

Differential Revision: https://reviews.llvm.org/D140961

19 months ago[flang] Initial support of allocate statement with source
Peixin Qiao [Fri, 13 Jan 2023 12:40:51 +0000 (20:40 +0800)]
[flang] Initial support of allocate statement with source

Support allocate statement with source in runtime version. The source
expression is evaluated only once for each allocate statement. When the
source expression has shape-spec, uses it for bounds. Otherwise, get
the bounds from the source expression. Get the length if the source
expression has deferred length parameter.

Reviewed By: clementval, jeanPerier

Differential Revision: https://reviews.llvm.org/D137812

19 months ago[lld][ARM] support position independent thunks for Armv4(T)
Ties Stuij [Fri, 13 Jan 2023 10:00:23 +0000 (10:00 +0000)]
[lld][ARM] support position independent thunks for Armv4(T)

- Position independent thunks now work for both Armv4 and Armv4T
- Armv4 arm->arm thunks don't emit a BX anymore, which doesn't exist for the
  arch. This fixes https://github.com/llvm/llvm-project/issues/50764.
- Armv4 and Armv4T both have the same arm->arm behaviour. Which also is
  desirable for the above ticket.

Reviewed By: MaskRay, peter.smith

Differential Revision: https://reviews.llvm.org/D141272

19 months ago[TableGen] Use raw_ostream::write_escaped instead of reinventing it. NFCI.
Jay Foad [Fri, 13 Jan 2023 11:12:17 +0000 (11:12 +0000)]
[TableGen] Use raw_ostream::write_escaped instead of reinventing it. NFCI.

19 months agoRecommit [SchedBoundary] Add dump method for resource usage.
Francesco Petrogalli [Fri, 13 Jan 2023 10:16:37 +0000 (11:16 +0100)]
Recommit [SchedBoundary] Add dump method for resource usage.

Summary:
As supporting information, I have added an example that describes how
the indexes of the vector of resources SchedBoundary::ReservedCycles
are tracked by the field SchedBoundary::ReservedCyclesIndex.

This has a minor rework of
https://github.com/llvm/llvm-project/commit/b39a9a94f420a25a239ae03097c255900cbd660e
which was reverted in
https://github.com/llvm/llvm-project/commit/df6ae1779fafd9984e144a27315d6dd65b32c325
becasue the llc invocation of the test was missing the argument
`-mtriple`.

See for example the failure at
https://lab.llvm.org/buildbot#builders/231/builds/7245 that reported
the following when targeting a non-aarch64 native build:

    'cortex-a55' is not a recognized processor for this target (ignoring processor)

Reviewers: jroelofs

Subscribers:

Differential Revision: https://reviews.llvm.org/D141367

19 months ago[SCEV] Support all Min/Max SCEVs for GetMinTrailingZeros
Joshua Cao [Thu, 12 Jan 2023 06:19:30 +0000 (22:19 -0800)]
[SCEV] Support all Min/Max SCEVs for GetMinTrailingZeros

There is already support for U/SMax. No reason why Min and SequentialMin
should not be supported.

NFC: code in GetMinTrailingZeroes is copied for a couple node types.
Refactor them into a single code block.

Differential Revision: https://reviews.llvm.org/D141568

19 months agoRevert "[mlir][SparseTensor] Add a few more tests for sparse vectorization"
Quentin Colombet [Fri, 13 Jan 2023 10:30:09 +0000 (10:30 +0000)]
Revert "[mlir][SparseTensor] Add a few more tests for sparse vectorization"

This reverts commit 904f2ccc3ba1d3aaf94140aa4595fd41af67d897.

This breaks a window bot. Reverting while I investigate.
https://lab.llvm.org/buildbot/#/builders/13/builds/30748

19 months ago[mlir][memref] Add details on the semantic of reinterpret_cast
Quentin Colombet [Fri, 13 Jan 2023 10:09:10 +0000 (10:09 +0000)]
[mlir][memref] Add details on the semantic of reinterpret_cast

Make it clearer what the semantic of reinterpret_cast is.

In particular, call out that this instruction is not a no-op.

NFC

Related to https://github.com/llvm/llvm-project/issues/59896

Differential Revision: https://reviews.llvm.org/D141662

19 months agoRevert "[SchedBoundary] Add dump method for resource usage."
Francesco Petrogalli [Fri, 13 Jan 2023 10:12:15 +0000 (11:12 +0100)]
Revert "[SchedBoundary] Add dump method for resource usage."

Reverting because of https://lab.llvm.org/buildbot#builders/16/builds/41860

When building on x86, I need to specify also -mtriple in the
invocation of llc otherwise the folllowing error shows up:

    'cortex-a55' is not a recognized processor for this target (ignoring processor)

This reverts commit b39a9a94f420a25a239ae03097c255900cbd660e.

19 months ago[flang] Implement codegen of hlfir.designate with component refs
Jean Perier [Fri, 13 Jan 2023 09:58:03 +0000 (10:58 +0100)]
[flang] Implement codegen of hlfir.designate with component refs

Lower all the different kinds of hlfir.designate component refs to
FIR.

Differential Revision: https://reviews.llvm.org/D141476

19 months ago[InstCombine] Regenerate test checks (NFC)
Nikita Popov [Fri, 13 Jan 2023 10:04:03 +0000 (11:04 +0100)]
[InstCombine] Regenerate test checks (NFC)

Pick up changes in variables names.

19 months ago[bazel] Updates for https://github.com/llvm/llvm-project/commit/aa0883b59ae17e5465906...
Dmitri Gribenko [Fri, 13 Jan 2023 09:55:26 +0000 (10:55 +0100)]
[bazel] Updates for https://github.com/llvm/llvm-project/commit/aa0883b59ae17e5465906dad51b5561b5292a28d

19 months ago[mlir] GreedyPatternRewriter: Add ancestors to worklist
Matthias Springer [Fri, 13 Jan 2023 09:42:01 +0000 (10:42 +0100)]
[mlir] GreedyPatternRewriter: Add ancestors to worklist

When adding an op to the worklist, also add its ancestors to the worklist. This allows for RewritePatterns to match an op `a` based on what is inside of the body of `a`.

This change fixes a problem that became apparent with `vector.warp_execute_on_lane_0`, but could probably be triggered with similar patterns. The pattern extracts an op `b` with `eligible = true` from the body of an op `a`:
```
test.a {
  %0 = test.b() {eligible = true}
  yield %0
}
```

Afterwards:
```
%0 = test.b() {eligible = true}
test.a {
  yield %0
}
```

The pattern is an `OpRewritePattern<OpA>`. For some reason, `test.a` is not on the GreedyPatternRewriter's worklist. E.g., because no pattern could be applied and it was removed. Now, another pattern updates `test.b`, so that `eligible` is changed from `true` to `false`. The `OpRewritePattern<OpA>` could now be applied, but (without this revision) `test.a` is still not on the worklist.

Note: In the above example, an `OpRewritePattern<OpB>` could have been used instead of an `OpRewritePattern<OpA>`. With such a design, we can run into the same problem (when the `eligible` attr is on `test.a` and `test.b` is removed from the worklist because no patterns could be applied).

Note: This change uncovered an unrelated bug in TestSCFUtils.cpp that was triggered due to a change in the order in which ops are processed. A TODO is added to the broken code and test cases are adapted so that the bug is no longer triggered.

Differential Revision: https://reviews.llvm.org/D140304

19 months ago[AMDGPU][AsmParser][NFC] Refine defining i8- and i16-typed custom operands.
Ivan Kosarev [Fri, 13 Jan 2023 09:33:11 +0000 (09:33 +0000)]
[AMDGPU][AsmParser][NFC] Refine defining i8- and i16-typed custom operands.

Reviewed By: dp

Differential Revision: https://reviews.llvm.org/D140799

19 months ago[mlir][SparseTensor] Add a few more tests for sparse vectorization
Quentin Colombet [Tue, 10 Jan 2023 10:58:01 +0000 (10:58 +0000)]
[mlir][SparseTensor] Add a few more tests for sparse vectorization

These tests cover muli, xor, and, addf, subf, and addi.

The tests themselves are not that interesting, their goal is to provide
code coverage for all the types of reductions currently supported.

NFC

Differential Revision: https://reviews.llvm.org/D141369

19 months ago[MemDep] Reduce block limit
Nikita Popov [Thu, 15 Dec 2022 13:05:00 +0000 (14:05 +0100)]
[MemDep] Reduce block limit

The non-local MemDep analysis has a limit on the number of blocks
it will scan trying to find dependencies. The current limit of 1000
is very high, especially when we consider that each block scan can
also visit up to 100 instructions. In degenerate cases (where we
actually scan that many blocks) MemDep/GVN dominate overall
compile-time, for little benefit.

This patch reduces the limit to 200, which is probably still too
large, but at least mitigates some of the more catastrophic cases.
(For comparison, MSSA clobber walks consider up to 100
MemoryDefs/MemoryPhis, rather than 200 blocks * 100 instructions,
but these limits aren't directly comparable.)

I know that we were kind of hoping that this issue would resolve
itself in time, either by a switch to NewGVN or use of MSSA in GVN.
But I think we should still address this in the meantime.
Additionally, a switch to an MSSA-based implementation will
effectively be doing this as well, in a roundabout way (by dint of
MSSA having lower cutoffs than MDA).

Differential Revision: https://reviews.llvm.org/D140097

19 months ago[SchedBoundary] Add dump method for resource usage.
Francesco Petrogalli [Fri, 13 Jan 2023 09:04:39 +0000 (10:04 +0100)]
[SchedBoundary] Add dump method for resource usage.

As supporting information, I have added an example that describes how
the indexes of the vector of resources SchedBoundary::ReservedCycles
are tracked by the field SchedBoundary::ReservedCyclesIndex.

Reviewed By: jroelofs

Differential Revision: https://reviews.llvm.org/D141367

19 months ago[Mips] Convert some tests to opaque pointers (NFC)
Nikita Popov [Thu, 12 Jan 2023 11:04:56 +0000 (12:04 +0100)]
[Mips] Convert some tests to opaque pointers (NFC)

I'm not sure why, but the absence of bitcasts / no-op GEPs causes
the branch delay slot to be used.

Differential Revision: https://reviews.llvm.org/D141593

19 months ago[mlir][affine][transform] Simplify affine.min/max ops with given constraints
Matthias Springer [Fri, 13 Jan 2023 09:23:51 +0000 (10:23 +0100)]
[mlir][affine][transform] Simplify affine.min/max ops with given constraints

This transform op uses `mlir::simplifyConstrainedMinMaxOp` to simplify `affine.min` and `affine.max` ops based on a given constraints.

Differential Revision: https://reviews.llvm.org/D140997

19 months ago[GVN][NFC] Refactor GVN::AnalyzeLoadAvailability method
Sergey Kachkov [Fri, 13 Jan 2023 08:21:55 +0000 (11:21 +0300)]
[GVN][NFC] Refactor GVN::AnalyzeLoadAvailability method

Simplify AnalyzeLoadAvailability code:
1. Use std::optional for return value
2. Use range-based loop for non-local dependencies

Differential Revision: https://reviews.llvm.org/D141664

19 months agoReland "[mlir][llvm] Add an explicit void type debug info attribute."
Tobias Gysi [Fri, 13 Jan 2023 09:14:21 +0000 (10:14 +0100)]
Reland "[mlir][llvm] Add an explicit void type debug info attribute."

Previously, the DISubroutineType attribute used an optional result
parameter and an optional argument types array to model the subroutine
signature. LLVM IR debug metadata, on the other hand, has one types
list whose first entry maps to the result type. That entry may be
null to model a void result type. The type list may also be entirely
empty not specifying any type information. The latter is problematic
since the current DISubroutineType attribute cannot express it.

The revision changes DISubroutineTypeAttr to closely follow the
LLVM metadata design. In particular, it uses a single types parameter
array to model the subroutine signature and introduces an explicit
DIVoidResultTypeAttr to model the null entries.

Reviewed By: Dinistro

Differential Revision: https://reviews.llvm.org/D141261

This reverts commit 81f57b6
and relands commit a960547

Fixes flang build and drop_begin on an empty array ref.

19 months ago[clang] Redefine some AVR specific macros
Ben Shi [Thu, 12 Jan 2023 11:58:53 +0000 (19:58 +0800)]
[clang] Redefine some AVR specific macros

Fixes https://github.com/llvm/llvm-project/issues/58855

Reviewed By: aykevl, Miss_Grape

Differential Revision: https://reviews.llvm.org/D141598

19 months ago[Driver] Add crtfastmath.o on Solaris if appropriate
Rainer Orth [Fri, 13 Jan 2023 09:08:33 +0000 (10:08 +0100)]
[Driver] Add crtfastmath.o on Solaris if appropriate

`Flang :: Driver/fast_math.f90` `FAIL`s on Solaris because `crtfastmath.o`
is missing from the link line.

This patch adds it as appropriate.

Tested on `amd64-pc-solaris2.11`.

Differential Revision: https://reviews.llvm.org/D141596

19 months agoMachineIRBuilder: Add buildMergeValues. NFC
Diana Picus [Mon, 9 Jan 2023 08:21:41 +0000 (09:21 +0100)]
MachineIRBuilder: Add buildMergeValues. NFC

Add a `buildMergeValues` method that unconditionally builds a
G_MERGE_VALUES instruction, as opposed to `buildMergeLikeInstr` which
may decide on a different opcode based on the input types.

I haven't audited all the uses of `buildMergeLikeInstr` to see if they
can be replaced with `buildMergeValues`, but I did find a couple of
obvious ones where we check that we're merging scalars right before
calling `buildMerge`.

This is a follow-up suggested in https://reviews.llvm.org/D140964

Differential Revision: https://reviews.llvm.org/D141373

19 months agoMachineIRBuilder: Rename buildMerge. NFC
Diana Picus [Mon, 9 Jan 2023 10:59:00 +0000 (11:59 +0100)]
MachineIRBuilder: Rename buildMerge. NFC

`buildMerge` may build a G_MERGE_VALUES, G_BUILD_VECTOR or
G_CONCAT_VECTORS. Rename it to `buildMergeLikeInstr`.

This is a follow-up suggested in https://reviews.llvm.org/D140964

Differential Revision: https://reviews.llvm.org/D141372

19 months agoGlobalISel: s/Op/Instr in some places. NFC
Diana Picus [Thu, 12 Jan 2023 10:01:54 +0000 (11:01 +0100)]
GlobalISel: s/Op/Instr in some places. NFC

This patch replaces `GMergeLikeOp` with `GMergeLikeInstr` and
`MachineIRBuilder::buildAssertOp` with `buildAssertInstr` in order to
remove ambiguity. Discussed in: https://reviews.llvm.org/D141372

19 months ago[flang] Lower elemental intrinsics to hlfir.elemental
Jean Perier [Fri, 13 Jan 2023 08:15:52 +0000 (09:15 +0100)]
[flang] Lower elemental intrinsics to hlfir.elemental

- Move the core code generating hlfir.elemental for user calls from
  genUserElementalCall into a new ElementalCallBuilder class and use
  C++ CRTP (curiously recursive template pattern) to implement the
  parts specific to user and intrinsic call into ElementalUserCallBuilder
  and ElementalIntrinsicCallBuilder. This allows sharing the core logic
  to lower elemental procedures for both user defined and intrinsics
  procedures.

- To allow using ElementalCallBuilder, split the intrinsic lowering code
  into two parts: first lower the arguments to hlfir::Entity regardless
  of the interface of the intrinsics, and then, in a different function
  (genIntrinsicProcRefCore), prepare the hlfir::Entity according to the
  interface. This allows using the same core logic to prepare "normal"
  arguments for non-elemental intrinsics, and to prepare the elements of
  array arguments inside elemental call (ElementalIntrinsicCallBuilder
  calls genIntrinsicProcRefCore once it has computed the scalar actual
  arguments).
  To allow this split, genExprBox/genExprAddr/genExprValue logic had to
  be split in ConvertExprToHlfir.[cpp/h].

- Add missing statement context pushScope/finalizeAndPop around the
  code generation inside the hlfir.elemental so that any temps created
  while lowering the call at the element level is correctly cleaned-up.

- One piece of code in hlfir::Entity::hasNonDefaultLowerBounds() was wrong for assumed shape arrays (returned true when an assumed shaped array had no explicit lower bounds). This caused the added test to hit a bogus TODO, so fix it.

Elemental intrinsics returning are still TODO (e.g., adjustl). I will implement this in a next patch, this one is big enough.

Differential Revision: https://reviews.llvm.org/D141612

19 months ago[libc][NFC] Use C headers in host CPU sniffing code.
Siva Chandra Reddy [Fri, 13 Jan 2023 07:42:49 +0000 (07:42 +0000)]
[libc][NFC] Use C headers in host CPU sniffing code.

19 months ago[C2x] reject type definitions in offsetof
Yingchi Long [Fri, 16 Sep 2022 14:13:54 +0000 (22:13 +0800)]
[C2x] reject type definitions in offsetof

https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2350.htm made very
clear that it is an UB having type definitions with in offsetof. After
this patch clang will reject any type definitions in __builtin_offsetof.

Fixes https://github.com/llvm/llvm-project/issues/57065

```
local/offsetof.c:10:38: error: 'struct S' cannot be defined in '__builtin_offsetof'
    return __builtin_offsetof(struct S{ int a, b;}, a);
                                     ^
```

Reviewed By: aaron.ballman, #clang-language-wg

Differential Revision: https://reviews.llvm.org/D133574

19 months ago[Instcombine] Add precommit tests for pattern xor(and, or); NFC
chenglin.bi [Fri, 13 Jan 2023 07:31:19 +0000 (15:31 +0800)]
[Instcombine] Add precommit tests for pattern xor(and, or); NFC

19 months ago[clang-tidy][doc] Deprecate the AnalyzeTemporaryDtors option
Carlos Galvez [Thu, 12 Jan 2023 10:10:48 +0000 (10:10 +0000)]
[clang-tidy][doc] Deprecate the AnalyzeTemporaryDtors option

It's not used anywhere, and we should not keep it
for eternity. Document it as deprecated and announce
its removal 2 releases later, so people have time
to update their .clang-tidy files.

Differential Revision: https://reviews.llvm.org/D141583

19 months ago[RISCV][VP] expand vp intrinsics if no +zve32x feature
Yingchi Long [Fri, 9 Dec 2022 13:05:37 +0000 (21:05 +0800)]
[RISCV][VP] expand vp intrinsics if no +zve32x feature

If the subtarget does not support VInstructions, expand vp intrinscs to scalar instructions.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D139706

19 months ago[AVR] Fix a bug in AsmPrinter when printing inline-asm operands
Ben Shi [Thu, 12 Jan 2023 09:21:55 +0000 (17:21 +0800)]
[AVR] Fix a bug in AsmPrinter when printing inline-asm operands

Fixes https://github.com/llvm/llvm-project/issues/58878

Reviewed By: aykevl, Miss_Grape

Differential Revision: https://reviews.llvm.org/D141589

19 months ago[X86] Remove unused variable after D140087
Vitaly Buka [Fri, 13 Jan 2023 05:46:51 +0000 (21:46 -0800)]
[X86] Remove unused variable after D140087

19 months ago[LoopReroll] Allow for multiple loop control only induction vars
Joshua Cao [Fri, 6 Jan 2023 04:48:13 +0000 (20:48 -0800)]
[LoopReroll] Allow for multiple loop control only induction vars

Before this, LoopReroll would fail an assertion, falsely assuming that
there can only possibly a single loop control only induction variable.

For example:

```
%a = phi i16 [ %dec2, %for.body ], [ 0, %entry ]
%b = phi i16 [ %dec1, %for.body ], [ 0, %entry ]
%a.next = add nsw i16 %1, -1
%b.next = add nsw i16 %0, -1
%add = add nsw i16 %a, %b
; ... rerollable code
%cmp.not = icmp eq i16 -10, %add
br i1 %cmp.not, label %exit, label %loop
```

Both %a and %b are valid loop control only induction vars

Additionally, some NFC changes to remove unnecessary isa<PHINode> check

Updated complex_reroll checks

Differential Revision: https://reviews.llvm.org/D141109

19 months ago[X86] Improve mul x, 2^N +/- 2 pattern by making the +/- 2x compute independently...
Noah Goldstein [Fri, 13 Jan 2023 02:30:00 +0000 (18:30 -0800)]
[X86] Improve mul x, 2^N +/- 2 pattern by making the +/- 2x compute independently to x << N

Previous pattern was omitting ops in sequence which just increases the
latency (to 3c, same as imul!) i.e:

`(add/sub (add/sub (shl x, N), x), x)`

Better is to compute 2x indepedently so x << N for better ULP i.e:
`(add/sub (shl x, N), (add x, x))`

Reviewed By: pengfei, RKSimon

Differential Revision: https://reviews.llvm.org/D141113

19 months ago[X86] Replace (31/63 -/^ X) with (NOT X) and ignore (32/64 ^ X) when computing shift...
Noah Goldstein [Fri, 13 Jan 2023 02:24:43 +0000 (18:24 -0800)]
[X86] Replace (31/63 -/^ X) with (NOT X) and ignore (32/64 ^ X) when computing shift count

Shift count is masked by hardware so these peepholes just extend
common patterns for NOT to the lower bits of shift count.

As well (32/64 ^ X) is masked off by the shift so can be safely
ignored.

Reviewed By: pengfei, lebedev.ri

Differential Revision: https://reviews.llvm.org/D140087

19 months agoAMDGPU/GlobalISel: Make regbankselect of implicit_def consistent with constants
Matt Arsenault [Fri, 30 Dec 2022 14:21:49 +0000 (09:21 -0500)]
AMDGPU/GlobalISel: Make regbankselect of implicit_def consistent with constants

19 months ago[RISCV] Change the return type of getStreamer() to support the use of overloading...
LiDongjin [Fri, 13 Jan 2023 03:48:22 +0000 (11:48 +0800)]
[RISCV] Change the return type of getStreamer() to support the use of overloading and other functions in RISCVELFStreamer

Move the declaration of RISCVELFStreamer from RISCVELFStreamer.cpp to RISCVELFStreamer.h.
Change the return type of getStreamer() to support the use of overloading and other functions in RISCVELFStreamer.

Differential Revision: https://reviews.llvm.org/D138500

19 months ago[lldb/test] Fix data racing issue in TestStackCoreScriptedProcess
Med Ismail Bennani [Wed, 11 Jan 2023 00:28:25 +0000 (16:28 -0800)]
[lldb/test] Fix data racing issue in TestStackCoreScriptedProcess

This patch should fix an nondeterministic error in TestStackCoreScriptedProcess.

In order to test both the multithreading capability and shared library
loading in Scripted Processes, the test would create multiple threads
that would take the same variable as a reference.

The first thread would alter the value and the second thread would
monitor the value until it gets altered. This assumed a certain ordering
regarding the `std::thread` spawning, however the ordering was not
always guaranteed at runtime.

To fix that, the test now makes use of a `std::condition_variable`
shared between the each thread. On the former, it will notify the other
thread when the variable gets initialized or updated and on the latter,
it will wait until the variable it receives a new notification.

This should fix the data racing issue while preserving the testing
coverage.

rdar://98678134

Differential Revision: https://reviews.llvm.org/D139484

Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
19 months ago[lldb] Update custom commands to always be overrriden
Med Ismail Bennani [Thu, 12 Jan 2023 23:30:24 +0000 (15:30 -0800)]
[lldb] Update custom commands to always be overrriden

This is a follow-up patch to 6f7835f309b9.

As explained previously, when running from an IDE, it can happen that
the IDE imports some lldb scripts by itself. If the user also tries to
import these commands, lldb will show the following message:

```
error: cannot add command: user command exists and force replace not set
```

This message is confusing to the user, because it suggests that the
command import failed and that the execution should stop. However, in
this case, lldb will continue the execution with the command added
previously by the user.

To prevent that, this patch updates every first-party lldb-packaged
custom commands to override commands that were pre-imported in lldb.

Differential Revision: https://reviews.llvm.org/D140293

Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
19 months ago[lldb] Force override when adding crashlog command
Med Ismail Bennani [Thu, 12 Jan 2023 23:29:19 +0000 (15:29 -0800)]
[lldb] Force override when adding crashlog command

When using interactive crashlog from an IDE, it can happen that the user
already have the `command script import lldb.macosx.crashlog` command on
their `lldbinit` file.

That leads to showing some message:

```
error: cannot add command: user command exists and force replace not set
error: cannot add command: user command exists and force replace not set
```

This leads to confusion because the crashlog symbolication continues and
succeeds even after these errors.

To address that, the crashlog commands get overridden everytime the
script get re-imported.

rdar://103403943

Differential Revision: https://reviews.llvm.org/D140113

Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
19 months ago[lldb/Process] Populate queues in Scripted Process
Med Ismail Bennani [Thu, 12 Jan 2023 23:27:45 +0000 (15:27 -0800)]
[lldb/Process] Populate queues in Scripted Process

This patch enhances queue support in Scripted Processes.

Scripted Threads could already report their queue name if they had one,
but this information was only surfaced when getting the process and
thread status.

However, no queue was create and added to the scripted process queue
list. This patch improves that by creating a queue from the scripted
thread queue name. For now, it uses an invalid queue id, since the
scripted thread doesn't expose this capability yet, but this could
easily be supported if the queue id information is available.

rdar://98844004

Differential Revision: https://reviews.llvm.org/D139853

Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
19 months ago[lldb/crashlog] Refactor CrashLogParser into a Factory pattern
Med Ismail Bennani [Thu, 12 Jan 2023 23:26:49 +0000 (15:26 -0800)]
[lldb/crashlog] Refactor CrashLogParser into a Factory pattern

This patch should fix an undefined behaviour that's happening when
parsing a crash report from an IDE. In the previous implementation, the
CrashLogParser base class would use the `__new__` static class method to
create the right parser instance depending on the crash report type.

For some reasons, the derived parser initializer wouldn't be called when
running the command from an IDE, so this patch refactors the
CrashLogParser code to replace the use of the `__new__` method with a
factory `create` static method.

rdar://100527640

Differential Revision: https://reviews.llvm.org/D139951

Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
19 months agoRemove a workaround for libstdc++4.8
Owen Anderson [Thu, 12 Jan 2023 04:53:30 +0000 (21:53 -0700)]
Remove a workaround for libstdc++4.8

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D141564

19 months ago[gn] Reformat all build files
Nico Weber [Fri, 13 Jan 2023 03:04:20 +0000 (22:04 -0500)]
[gn] Reformat all build files

Ran:

    git ls-files '*.gn' '*.gni' | xargs llvm/utils/gn/gn.py format

19 months ago[LinkerWrapper] Adjust test after a previous change in temp file names
Joseph Huber [Fri, 13 Jan 2023 02:35:51 +0000 (20:35 -0600)]
[LinkerWrapper] Adjust test after a previous change in temp file names

Summary:
This patch adjusts the test to correctly search for the temp files
created after a previous patch changed the name. This wasn't caught
because it still technically worked.

19 months ago[clangd] support expanding `decltype(expr)`
v1nh1shungry [Fri, 13 Jan 2023 02:15:13 +0000 (03:15 +0100)]
[clangd] support expanding `decltype(expr)`

Enable the existing tweak `ExpandAutoType` to expand
`decltype(expr)`, e.g.

```
decltype(0) i;
```

will expand to

```
int i;
```

Therefore, rename the tweak `ExpandAutoType` to `ExpandDeducedType`.

This patch also fixes some nits,

* avoid replacing reference to a function
* avoid replacing array types and reference to an array

Fixes https://github.com/clangd/clangd/issues/1456

Reviewed By: sammccall

Differential Revision: https://reviews.llvm.org/D141226

19 months ago[LLDB][LoongArch] Add unittests for EmulateInstructionLoongArch
Hui Li [Fri, 13 Jan 2023 02:17:48 +0000 (10:17 +0800)]
[LLDB][LoongArch] Add unittests for EmulateInstructionLoongArch

Add unit tests For EmulateInstructionLoongArch existing branch instruction.
Add 19 test cases in total.

Without this patch:

```
$ ninja check-lldb-unit
[0/1] Running lldb unit test suite

Testing Time: 10.55s
  Passed: 1025
```

With this patch:

```
$ ninja check-lldb-unit
[0/1] Running lldb unit test suite

Testing Time: 10.45s
  Passed: 1044
```

Reviewed By: DavidSpickett

Differential Revision: https://reviews.llvm.org/D140386

19 months ago[LLDB][LoongArch] ObjectFile: add a case for `EM_LOONGARCH`
Hui Li [Fri, 13 Jan 2023 02:16:29 +0000 (10:16 +0800)]
[LLDB][LoongArch] ObjectFile: add a case for `EM_LOONGARCH`

This adds the jump slot mapping for LoongArch. This is a simple
change that ensures lldb running properly.

 Note that this problem was found only when CMake variables
 "DLLVM_ENABLE_ASSERTIONS=ON" is selected.

Without this patch,

```
$ build/bin/lldb  test

../ELFHeader::GetRelocationJumpSlotType() const: Assertion `false && "architecture not supported"' fail

```
With this patch

```
$ build/bin/lldb  test
(lldb) target create "test"
Current executable set to '../test' (loongarch64).

```

Reviewed By: SixWeining, DavidSpickett

Differential Revision: https://reviews.llvm.org/D141245

19 months agoRevert "[MergeICmps] Adapt to non-eq comparisons"
Vitaly Buka [Fri, 13 Jan 2023 02:09:15 +0000 (18:09 -0800)]
Revert "[MergeICmps] Adapt to non-eq comparisons"

Breaks ubsan build, details in D141188.

This reverts commit 3ac2b3a4f9effc9f79822e770f209fd70ff66362.

19 months ago[AMDGPU][NFC] Heterogeneous DWARF extensions update
Tony Tye [Thu, 5 Jan 2023 00:34:02 +0000 (00:34 +0000)]
[AMDGPU][NFC] Heterogeneous DWARF extensions update

- Clarify CFI rules in heterogeneous DWARF extensions

- Added DWARF source language memory spaces.

- Added DWARF architecture address spaces.

- Other minor corrections.

Reviewed By: scott.linder

Differential Revision: https://reviews.llvm.org/D141548

19 months ago[clang][test] Remove unnecessary 'REQUIRES'
Ben Shi [Fri, 13 Jan 2023 01:25:42 +0000 (09:25 +0800)]
[clang][test] Remove unnecessary 'REQUIRES'

The test 'Preprocessor/predefined-arch-macros.c' contains many
target tests than 'amdgpu'. If clang is built without 'amdgpu',
then failures in other target tests will not be reported.

Reviewed By: aaron.ballman, MaskRay

Differential Revision: https://reviews.llvm.org/D141647

19 months ago[LSAN] More LSAN interface tweaking.
Kirill Stoimenov [Thu, 12 Jan 2023 21:31:49 +0000 (13:31 -0800)]
[LSAN] More LSAN interface tweaking.

Main goal is to remove thread registry dependency from the interface because HWASAN is using its own code to manage threads.

Reviewed By: vitalybuka, kstoimenov

Differential Revision: https://reviews.llvm.org/D140039

19 months ago[LinkerWrapper] Fix test on Windows
Joseph Huber [Fri, 13 Jan 2023 01:56:24 +0000 (19:56 -0600)]
[LinkerWrapper] Fix test on Windows

Summary:
`clang` is called `clang.exe` on windows, we need to glob the text
inbetween.

19 months agoRemove a bogus assertion.
Richard Smith [Fri, 13 Jan 2023 01:21:45 +0000 (17:21 -0800)]
Remove a bogus assertion.

The result of recomputing the type of an LValue may be different from
the type we find serialized with that LValue, either due to differing
type sugar in a redeclaration, or more significantly due to a later
declaration of a variable having an array bound that a prior declaration
did not have. Therefore this assertion was attempting to verify a
property that isn't necessarily true.

19 months ago[lld/mac] Add support for distributed ThinLTO
Nico Weber [Sun, 20 Nov 2022 16:59:16 +0000 (11:59 -0500)]
[lld/mac] Add support for distributed ThinLTO

Adds support for the following flags:

* --thinlto-index-only, --thinlto-index-only=
* --thinlto-emit-imports-files
* --thinlto-emit-index-files
* --thinlto-object-suffix-replace=
* --thinlto-prefix-replace=

See https://blog.llvm.org/2016/06/thinlto-scalable-and-incremental-lto.html
for some words on --thinlto-index-only.

I don't really need the other flags, but they were in the vicinity
and _someone_ might need them, so I figured I'd add them too.

`-object_path_lto` now sets `c.AlwaysEmitRegularLTOObj` as in the other ports,
which means it can now only point to a filename for non-thin LTO.
I think that was the intent of D129705 anyways, so update
test/MachO/lto-object-path.ll to use a non-thin bitcode file for that test.

Differential Revision: https://reviews.llvm.org/D138451

19 months ago[MemProf] Collect access density statistics during profiling
Teresa Johnson [Thu, 12 Jan 2023 01:16:03 +0000 (17:16 -0800)]
[MemProf] Collect access density statistics during profiling

Track min/max/avg access density (accesses per byte and accesses per
byte per lifetime second) metrics directly during profiling. This allows
more accurate use of these metrics in profile analysis and use, instead
of trying to compute them from already aggregated data in the profile.

This required regenerating some of the raw profile and executable inputs
for a few tests. While here, make the llvm-profdata memprof tests more
resilient to differences in things like memory mapping, timestamps and
cpu ids to make future test updates easier.

Differential Revision: https://reviews.llvm.org/D141558

19 months ago[RISCV] Custom lower vector ISD::CTLZ to the minimum of ISD::CTLZ_ZERO_UNDEF and...
Yeting Kuo [Thu, 12 Jan 2023 10:14:24 +0000 (18:14 +0800)]
[RISCV] Custom lower vector ISD::CTLZ to the minimum of ISD::CTLZ_ZERO_UNDEF and element size.

D111904 made RISC-V customized lower ISD::CTLZ_ZERO_UNDEF by converting to float
and using the float result. The expected value of CTLZ with zero input is the
element size of input type. Since the result of above method with zero input
must be greater than the element size, for ISD::CTLZ, we could use the minimum
of element size and the result of CTLZ_ZERO_UNDER with same input.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D141585

19 months agoRevert "[LSAN] More LSAN interface tweaking."
Vitaly Buka [Fri, 13 Jan 2023 01:30:26 +0000 (17:30 -0800)]
Revert "[LSAN] More LSAN interface tweaking."

Breaks bots.
Also it's missing changes we discussed on review.

This reverts commit f001e50f955c3cdf2deb79e38a9fd19c9a781882.
This reverts commit 2924189233fdb724453ead4b94595107b1ce9cfa.

19 months ago[Clang][RISCV] Update operand order for vmerge and vcompress
eopXD [Tue, 27 Dec 2022 09:12:51 +0000 (01:12 -0800)]
[Clang][RISCV] Update operand order for vmerge and vcompress

From:
  vint32m1_t vmerge_vvm_i32m1 (vbool32_t mask, vint32m1_t op1, vint32m1_t op2, size_t vl);
  vint32m1_t vcompress_vm_i32m1 (vbool32_t mask, vint32m1_t src, size_t vl);

To:
  vint32m1_t vmerge_vvm_i32m1 (vint32m1_t op1, vint32m1_t op2, vbool32_t selector, size_t vl);
  vint32m1_t vcompress_vm_i32m1 (vint32m1_t src, vbool32_t selector, size_t vl);

Address issues:
riscv-non-isa/rvv-intrinsic-doc#140
riscv-non-isa/rvv-intrinsic-doc#167

Pull request:
riscv-non-isa/rvv-intrinsic-doc#185

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D140686

19 months ago[gn] port 3432f4bf86e7 (//llvm/lib/Testing/Annotations)
Nico Weber [Fri, 13 Jan 2023 01:07:55 +0000 (20:07 -0500)]
[gn] port 3432f4bf86e7 (//llvm/lib/Testing/Annotations)

19 months ago[clang-tidy] don't warn when returning the result for bugprone-standalone-empty
v1nh1shungry [Thu, 12 Jan 2023 23:40:05 +0000 (23:40 +0000)]
[clang-tidy] don't warn when returning the result for bugprone-standalone-empty

Relevant issue: https://github.com/llvm/llvm-project/issues/59517

Currently this check will warn when the result is used in a `return`
statement, e.g.

```
bool foobar() {
  std::vector<int> v;
  return v.empty();
  // will get a warning here, which makes no sense IMO
}
```

Reviewed By: cjdb, denik

Differential Revision: https://reviews.llvm.org/D141107

19 months ago[lld-macho] Downgrade missing fat arch to warning
Keith Smiley [Thu, 12 Jan 2023 22:43:12 +0000 (14:43 -0800)]
[lld-macho] Downgrade missing fat arch to warning

This mirrors ld64's behavior. In many cases this likely still leads to a
link failure but if you didn't actually use anything from the library
it can be ignored. If you care about these invalid cases -fatal_warnings
still upgrades it back to an error.

Example: https://github.com/keith/ld64.lld/issues/3

Differential Revision: https://reviews.llvm.org/D141638

19 months ago[lsan] Fix build for Fuchsia after D140039
Alex Brachet [Fri, 13 Jan 2023 00:47:46 +0000 (00:47 +0000)]
[lsan] Fix build for Fuchsia after D140039

19 months agoAdd -disable-check-noreturn-call for stack protect
Xiang1 Zhang [Wed, 11 Jan 2023 08:54:07 +0000 (16:54 +0800)]
Add -disable-check-noreturn-call for stack protect

Differential Revision: https://reviews.llvm.org/D141556
Reviewed By: lanza

19 months agoOpenMP: Use inbounds in EmitOMPAggregateAssign
Matt Arsenault [Thu, 12 Jan 2023 01:34:00 +0000 (20:34 -0500)]
OpenMP: Use inbounds in EmitOMPAggregateAssign

This looked like a plausibly correct out of tree patch.
The changed testcases with the pragmas stripped out only use
inbounds GEPs so I assume this is correct.

19 months agoRevert "[GWP-ASan] Add recoverable mode."
Mitch Phillips [Thu, 12 Jan 2023 23:58:36 +0000 (15:58 -0800)]
Revert "[GWP-ASan] Add recoverable mode."

This reverts commit 90a9beb7cc9755791caa23dfc4e36bc544e98ed3.

Reason: Broke an internal build.

19 months agoRevert "[GWP-ASan] Fix test to work with Fuchsia's zxtest"
Mitch Phillips [Thu, 12 Jan 2023 23:58:10 +0000 (15:58 -0800)]
Revert "[GWP-ASan] Fix test to work with Fuchsia's zxtest"

This reverts commit 64138725e4cf0ab5e3f9c0dba38e11ce3d3b4017.

Parent change https://reviews.llvm.org/D140173 broke an internal build.
I'll land this bit as part of the re-land for
https://reviews.llvm.org/D140173.

19 months agoRevert "[GWP-ASan] Fix 6413872"
Mitch Phillips [Thu, 12 Jan 2023 23:57:27 +0000 (15:57 -0800)]
Revert "[GWP-ASan] Fix 6413872"

This reverts commit df87e62cbd4b5090b5349247d29457706d2ac4e8.

Parent change https://reviews.llvm.org/D140173 broke an internal build.
I'll integrate this patch into the re-land for
https://reviews.llvm.org/D140173.

19 months ago[LSAN] More LSAN interface tweaking.
Kirill Stoimenov [Thu, 12 Jan 2023 00:03:41 +0000 (00:03 +0000)]
[LSAN] More LSAN interface tweaking.

Main goal is to remove thread registry dependency from the interface because HWASAN is using its own code to manage threads.

Reviewed By: vitalybuka, kstoimenov

Differential Revision: https://reviews.llvm.org/D140039

19 months ago[OpenMP][FIX] The thread limit is the block, not grid size
Johannes Doerfert [Thu, 12 Jan 2023 23:29:55 +0000 (15:29 -0800)]
[OpenMP][FIX] The thread limit is the block, not grid size

19 months ago[CUDA] added cmath wrappers to unbreak CUDA compilation after D79555
Artem Belevich [Thu, 12 Jan 2023 01:09:21 +0000 (17:09 -0800)]
[CUDA] added cmath wrappers to unbreak CUDA compilation after D79555

libc++ introduced a handful of internal functions that may or may not be
constexpr, depending on C++ version. For pre-constexpr variants we must declare
__device__ counterparts. Otherwise the code fails to compile on the GPU side.
See https://reviews.llvm.org/D79555

Differential Revision: https://reviews.llvm.org/D141555

19 months ago[mlir][linalg] Fix incorrect reduction detection in Vectorizer
Diego Caballero [Thu, 12 Jan 2023 23:11:45 +0000 (23:11 +0000)]
[mlir][linalg] Fix incorrect reduction detection in Vectorizer

When detecting reductions, make sure the block argument is from the linalg generic op.
This fixes https://github.com/iree-org/iree/issues/11779.

Co-authored-by: Andrzej Warzynski <andrzej.warzynski@arm.com>
Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D141413

19 months ago[mlir][sparse] Minor code clean up (NFC).
bixia1 [Thu, 12 Jan 2023 00:25:50 +0000 (16:25 -0800)]
[mlir][sparse] Minor code clean up (NFC).

Reviewed By: Peiming

Differential Revision: https://reviews.llvm.org/D141554

19 months ago[Attributor][FIX] Consistently use the access/remote instruction
Johannes Doerfert [Thu, 12 Jan 2023 23:19:35 +0000 (15:19 -0800)]
[Attributor][FIX] Consistently use the access/remote instruction

We use the actual access (=remote) instruction when reasoning about
accesses, except for one leftover use case of the local instruction.
This caused us to potentially ignore the dominating write if the read
and write were in a different function than the (stack) allocation.

Reported by @ye-luo

19 months ago[libc++] Add clang-tidy to the list of possible substitutions for %{clang-tidy}
Nikolas Klauser [Mon, 9 Jan 2023 00:39:58 +0000 (01:39 +0100)]
[libc++] Add clang-tidy to the list of possible substitutions for %{clang-tidy}

Most people don't have a versioned clang-tidy locally. This allows them to still run the clang-tidy test if they have a suitable version.

Reviewed By: ldionne, #libc

Spies: libcxx-commits, arichardson

Differential Revision: https://reviews.llvm.org/D141241

19 months ago[mlir][NFC] Use explicit ArrayRef construction in OptionalArrayRefParameter
River Riddle [Thu, 12 Jan 2023 22:43:32 +0000 (14:43 -0800)]
[mlir][NFC] Use explicit ArrayRef construction in OptionalArrayRefParameter

GCC was barfing on template deduction for ArrayRef parameters in
some situations.

19 months ago[mlir][AttrType] Emit unbalanced character errors using the last punctuation used
River Riddle [Thu, 12 Jan 2023 22:20:00 +0000 (14:20 -0800)]
[mlir][AttrType] Emit unbalanced character errors using the last punctuation used

This gives a better diagnostic in general, because it indicates that the user didn't
close out the last open puncutation range. For example:

```
foo.op {
  some.op -> !blah.pointer<
}
```
We want the error to hint about the unclosed `<`, not the `}` (which isn't really in the context of the type).

Differential Revision: https://reviews.llvm.org/D141635

19 months ago[NFC] [scudo] avoid casting in a circle
Florian Mayer [Thu, 12 Jan 2023 22:47:03 +0000 (14:47 -0800)]
[NFC] [scudo] avoid casting in a circle

19 months ago[libTooling] Add `getFileRange` as an alternative to `getRangeForEdit`
Eric Li [Thu, 12 Jan 2023 22:10:39 +0000 (17:10 -0500)]
[libTooling] Add `getFileRange` as an alternative to `getRangeForEdit`

Add a `getFileRange` function alongside the existing `getRangeForEdit`
as a way to get a contiguous range within a single file (similar to
`getRangeForEdit`) but without the restriction that it cannot be in a
system header.

This can be used where a tool may want to use the range to extract the
source text. In such cases, we don't want to restrict this from
pulling from system headers.

Differential Revision: https://reviews.llvm.org/D141634

19 months ago[LinkerWrapper] Improve -save-temps behaviour
Joseph Huber [Thu, 12 Jan 2023 22:42:30 +0000 (16:42 -0600)]
[LinkerWrapper] Improve -save-temps behaviour

Summary:
This patch changes around some of the names for the temp files we
generate. This also emits the `postlink` and `postopt` temp files. We
don't use the built-in save temps from LTO right now because it doesn't
give the ASM properly. We may be able to fix that in the future.

19 months ago[AArch64] Add IfConv test with loads with invariant ops as conditions.
Florian Hahn [Thu, 12 Jan 2023 22:38:39 +0000 (22:38 +0000)]
[AArch64] Add IfConv test with loads with invariant ops as conditions.

Add EarlyIfConversion tests for AArch64 where the condition is a load
with invariant operands. A follow-up patch will propose a heuristic to
consider such conditions predictable.

19 months ago[mlir][sparse] support parsing slices in sparse tensor encoding attribute
Peiming Liu [Tue, 27 Dec 2022 22:10:37 +0000 (22:10 +0000)]
[mlir][sparse] support parsing slices in sparse tensor encoding attribute

Reviewed By: bixia

Differential Revision: https://reviews.llvm.org/D140712

19 months ago[libc][obvious] Add a missing dep.
Siva Chandra Reddy [Thu, 12 Jan 2023 22:33:38 +0000 (22:33 +0000)]
[libc][obvious] Add a missing dep.

19 months agoRevert "[SimplifyCFG] When eliminating `unreachable` landing pads, mark `call`s as...
Roman Lebedev [Thu, 12 Jan 2023 22:12:07 +0000 (01:12 +0300)]
Revert "[SimplifyCFG] When eliminating `unreachable` landing pads, mark `call`s as `nounwind`"

The bool is in the wrong place and might get implicitly converted from
the previous second argument - a pointer. Thinking about it more,
it's not really the best place for that functionality anyways,
only a single caller needs that.

This reverts commit 3c5b1f2d94d021005ce3769a4402d4a4ae843989.

19 months ago[mlir][python] fix python build
Jeff Niu [Thu, 12 Jan 2023 22:17:32 +0000 (14:17 -0800)]
[mlir][python] fix python build

19 months ago[lldb/test] Disable TestScriptedProcess.py on linux while investigating the issue
Med Ismail Bennani [Thu, 12 Jan 2023 22:10:48 +0000 (14:10 -0800)]
[lldb/test] Disable TestScriptedProcess.py on linux while investigating the issue

Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
19 months ago[SimplifyCFG] When eliminating `unreachable` landing pads, mark `call`s as `nounwind`
Roman Lebedev [Thu, 12 Jan 2023 21:05:19 +0000 (00:05 +0300)]
[SimplifyCFG] When eliminating `unreachable` landing pads, mark `call`s as `nounwind`

We really can't recover that knowledge, and `nounwind` knowledge,
(and not just a lack of the unwind edge, aka `call` instead of `invoke`),
is e.g. part of the reasoning in e.g. `mayHaveSideEffects()`.

Note that this is call-site-specific knowledge,
just because some callsite had an `unreachable`
unwind edge, does not mean that all will.

19 months ago[NFC][SimplifyCFG] Autogenerate checklines in some tests that eliminate unwind edges
Roman Lebedev [Thu, 12 Jan 2023 20:43:38 +0000 (23:43 +0300)]
[NFC][SimplifyCFG] Autogenerate checklines in some tests that eliminate unwind edges

19 months ago[test] Split out Annotations from `TestingSupport`
Jordan Rupprecht [Thu, 12 Jan 2023 21:40:47 +0000 (13:40 -0800)]
[test] Split out Annotations from `TestingSupport`

The Annotations helper class does not have a gtest or gmock dependency, but because it's bundled with the rest of TestingSupport, it gets one. By splitting it out, a target can use it without being forced to use LLVM's copy of gtest.

Reviewed By: GMNGeoffrey, sammccall, gribozavr2

Differential Revision: https://reviews.llvm.org/D141175

19 months ago[ORC] Tidy up MachOPlatform's references to ORC runtime registration functions.
Lang Hames [Thu, 12 Jan 2023 16:32:58 +0000 (08:32 -0800)]
[ORC] Tidy up MachOPlatform's references to ORC runtime registration functions.

Use an private struct, RuntimeFunction, to to keep the name and address of each
registration function together, and rename the member variables with their purpose
rather than the full name of the function in the runtime.

19 months ago[mlir][ods] Generate inferReturnTypes for ops with TypesMatchWith
Jeff Niu [Sun, 8 Jan 2023 20:46:52 +0000 (12:46 -0800)]
[mlir][ods] Generate inferReturnTypes for ops with TypesMatchWith

Ops that use TypesMatchWith to constrain result types for verification
and to infer result types during parser generation should also be able
to have the `inferReturnTypes` method auto generated. This patch
upgrades the logic for generating `inferReturnTypes` to handle the
TypesMatchWith trait by building a type inference graph where each edge
corresponds to "type of A can be inferred from type of B", supporting
transformers other than `"$_self"`.

Reviewed By: lattner, rriddle

Differential Revision: https://reviews.llvm.org/D141231

19 months ago[BOLT] Minor refactoring for -print-sorted-by option
Maksim Panchenko [Tue, 10 Jan 2023 03:14:38 +0000 (19:14 -0800)]
[BOLT] Minor refactoring for -print-sorted-by option

Only display used values for -print-sorted-by option when printing help.

Differential Revision: https://reviews.llvm.org/D141209

19 months ago[mlir] Add operations to BlockAndValueMapping and rename it to IRMapping
Jeff Niu [Sun, 8 Jan 2023 22:15:07 +0000 (14:15 -0800)]
[mlir] Add operations to BlockAndValueMapping and rename it to IRMapping

The patch adds operations to `BlockAndValueMapping` and renames it to `IRMapping`. When operations are cloned, old operations are mapped to the cloned operations. This allows mapping from an operation to a cloned operation. Example:

```
Operation *opWithRegion = ...
Operation *opInsideRegion = &opWithRegion->front().front();

IRMapping map
Operation *newOpWithRegion = opWithRegion->clone(map);
Operation *newOpInsideRegion = map.lookupOrNull(opInsideRegion);
```

Migration instructions:
All includes to `mlir/IR/BlockAndValueMapping.h` should be replaced with `mlir/IR/IRMapping.h`. All uses of `BlockAndValueMapping` need to be renamed to `IRMapping`.

Reviewed By: rriddle, mehdi_amini

Differential Revision: https://reviews.llvm.org/D139665