platform/upstream/llvm.git
21 months ago[NFC] Refactor dxil metadata code
Chris Bieneman [Wed, 21 Sep 2022 23:12:43 +0000 (18:12 -0500)]
[NFC] Refactor dxil metadata code

DXIL relies on a whole bunch of IR metadata constructs being populated
in the right shape. Rather than just hard coding or using complicated
arrangements of constant data strings, let's make first-class objects
that reprensent the metadata and manage reading and writing the
metadata from the module.

Reviewed By: python3kgae

Differential Revision: https://reviews.llvm.org/D134397

21 months ago[gn build] port 47afaf2eb02b (exegesis all targets)
Nico Weber [Thu, 22 Sep 2022 17:16:49 +0000 (13:16 -0400)]
[gn build] port 47afaf2eb02b (exegesis all targets)

21 months ago[flang][runtime] Fixes for element size calculation.
Slava Zakharin [Thu, 22 Sep 2022 16:31:20 +0000 (09:31 -0700)]
[flang][runtime] Fixes for element size calculation.

BytesFor() used to return KIND for the size, which is not always
correct, so I changed it to return the size of the actual CppType
corresponding to the given category and kind.

MinElemLen() used to calculate size incorrectly (e.g. CFI_type_extended_double
was sized 10, whereas it may occupy more bytes on a target), so I changed it
to call BytesFor().

Additional changes were needed to resolve new failures for transformational
intrinsics. These intrinsics used to work for not fully supported
data types (e.g. REAL(3)), but now stopped working because CppType
cannot be computed for those categories/kinds. The solution is to use
known element size from the source argument(s) for establishing
the destination descriptor - the element size is all that is needed
for transformational intrinsics to keep working.

Note that this does not help cases, where runtime still has to
compute the element size, e.g. when it creates descriptors for
components of derived types. If the component has unsupported
data type, BytesFor() will still fail. So these cases require
adding support for the missing types.

New regression unit test in Runtime/Transformational.cpp
demonstrates the case that will start working properly with
this commit.

21 months ago[clang-tidy] Add option WarnOnSizeOfPointerToAggregate.
Michael Benfield [Wed, 21 Sep 2022 19:46:35 +0000 (19:46 +0000)]
[clang-tidy] Add option WarnOnSizeOfPointerToAggregate.

This is now an option under the check bugprone-sizeof-expression.

Differential Revision: https://reviews.llvm.org/D134381

21 months ago[clang][DebugInfo] Emit access specifiers for typedefs
Jonathan Camilleri [Thu, 22 Sep 2022 16:59:00 +0000 (16:59 +0000)]
[clang][DebugInfo] Emit access specifiers for typedefs

The accessibility level of a typedef or using declaration in a
struct or class was being lost when producing debug information.

Reviewed By: dblaikie

Differential Revision: https://reviews.llvm.org/D134339

21 months ago[DebugInfo] Emit access specifiers for typedefs
Jonathan Camilleri [Thu, 22 Sep 2022 16:47:02 +0000 (16:47 +0000)]
[DebugInfo] Emit access specifiers for typedefs

The accessibility level of a typedef or using declaration in a
struct or class was being lost when producing debug information.

Reviewed By: dblaikie

Differential Revision: https://reviews.llvm.org/D134339

21 months agoUpdate the C status page for WG14 N2359
Aaron Ballman [Thu, 22 Sep 2022 17:04:27 +0000 (13:04 -0400)]
Update the C status page for WG14 N2359

We don't yet implement this paper, so this adds a basic test
demonstrating that.

21 months ago[mlir][vector] Fix test for vector-warp-distribute
Christopher Bate [Thu, 22 Sep 2022 16:11:59 +0000 (10:11 -0600)]
[mlir][vector] Fix test for vector-warp-distribute

Test does a "CHECK-NOT" against the function name when it should check
to ensure that the `vector.warp_execute_on_lane_0` is removed.

Reviewed By: ThomasRaoux

Differential Revision: https://reviews.llvm.org/D134448

21 months ago[AArh64-SVE]: Improve cost model for div/udiv/mul 128-bit vector operations
Hassnaa Hamdi [Tue, 23 Aug 2022 15:22:52 +0000 (15:22 +0000)]
[AArh64-SVE]: Improve cost model for div/udiv/mul 128-bit vector operations

Differential Revision: https://reviews.llvm.org/D132477

21 months ago[Doc][OpenCL] Fixed typos in code examples
Anastasia Stulova [Thu, 22 Sep 2022 16:46:47 +0000 (17:46 +0100)]
[Doc][OpenCL] Fixed typos in code examples

21 months ago[CostModel][X86] Remove duplicate ashr v4i64 cost table entry. NFCI.
Simon Pilgrim [Thu, 22 Sep 2022 16:27:20 +0000 (17:27 +0100)]
[CostModel][X86] Remove duplicate ashr v4i64 cost table entry. NFCI.

21 months agoMC: make section classification a bit more thorough
Saleem Abdulrasool [Fri, 5 Aug 2022 16:52:03 +0000 (16:52 +0000)]
MC: make section classification a bit more thorough

This does *NOT* change the emitted section flags in any way.  This only
impacts the internal classification of sections.

Extend the section classification in LLVM for ELF targets.  This has one
important change: we now classify sections as text by default rather
than readonly.  This matches the behaviour for GAS better.

Ensure that any section that has a writable attribute set is not treated
as readonly.  We also special case any section named `.debug_` which is
reserved for DWARF as metadata.  In the case none of the attributes are
set (or because no attributes were provided), consult the section name
for classification.  We match the well known names and classify the
section accordingly.  Any remaining section is now classified as text.

This change allows us to classify sections in the MC layer more
precisely which is needed for subsequent changes for handling target
specific behaviour.

Re-apply the change that was reverted with additional changes to
classify section prefixes appropriately and differentiate the TLS
sections, addressing the FIXME and post-commit review comments by
@MaskRay.

Differential Revision: https://reviews.llvm.org/D133456
Reviewed By: @MaskRay

21 months ago[mlir][sparse] Add codegen rule for the push_back operator.
bixia1 [Wed, 21 Sep 2022 16:26:09 +0000 (09:26 -0700)]
[mlir][sparse] Add codegen rule for the push_back operator.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D134372

21 months ago[llvm-exegesis] Initialize all supported targets
Philip Reames [Thu, 22 Sep 2022 16:02:22 +0000 (09:02 -0700)]
[llvm-exegesis] Initialize all supported targets

Enable registration of multiple exegesis targets at once. Use idiomatic approach to defining target select macros, but leave code in the llvm-mca sub-directories for now.

This is a stepping stone towards allowing llvm-exegesis benchmarking via simulator or testing in non-target dependent tests.

Differential Revision: https://reviews.llvm.org/D133605

21 months ago[VPlan] Only generate single instr for unpredicated stores of varying value to invari...
Philip Reames [Thu, 22 Sep 2022 15:47:19 +0000 (08:47 -0700)]
[VPlan] Only generate single instr for unpredicated stores of varying value to invariant address

This extends the previously added uniform store case to handle stores of loop varying values to a loop invariant address. Note that the placement of this code only allows unpredicated stores; this is important for correctness. (That is "IsPredicated" is always false at this point in the function.)

This patch does not include scalable types. The diff felt "large enough" as it were; I'll handle that in a separate patch. (It requires some changes to cost modeling.)

Differential Revision: https://reviews.llvm.org/D133580

21 months ago[RISCV] Use structured bindings in common RVV lowering code
Fraser Cormack [Thu, 22 Sep 2022 15:07:27 +0000 (16:07 +0100)]
[RISCV] Use structured bindings in common RVV lowering code

This patch uses structured bindings to simplify a couple of specific
cases when lowering RVV operations where we commonly declare two
SDValues and immediately 'tie' them to the mask and vector length.

There's also a couple places where we split vectors that structured
bindings make sense to use.

This patch tries to keep these sorts of changes minimal and to cases
where the returned types are commonly understood, rather than applying
this wholesale to the RISCV backend.

Reviewed By: reames

Differential Revision: https://reviews.llvm.org/D134442

21 months ago[AArch64] add tests for vector load combining; NFC
Sanjay Patel [Thu, 22 Sep 2022 15:21:08 +0000 (11:21 -0400)]
[AArch64] add tests for vector load combining; NFC

More coverage for D133584

21 months ago[CostModel][X86] Add CostKinds test coverage for mul-by-constant patterns
Simon Pilgrim [Thu, 22 Sep 2022 15:40:50 +0000 (16:40 +0100)]
[CostModel][X86] Add CostKinds test coverage for mul-by-constant patterns

Help check to see the costs predicted for mul->shift conversions

21 months ago[RISCV] Verify consistency of a couple TSFlags related to vector operands
Philip Reames [Thu, 22 Sep 2022 15:27:23 +0000 (08:27 -0700)]
[RISCV] Verify consistency of a couple TSFlags related to vector operands

Various bits of existing code assume the presence of one operand implies the presence of another.  Add verifier rules to catch violations.

Differential Revision: https://reviews.llvm.org/D133810

21 months ago[RISCV] Improve support for vector fp_to_sint_sat/uint_sat.
Craig Topper [Thu, 22 Sep 2022 15:13:47 +0000 (08:13 -0700)]
[RISCV] Improve support for vector fp_to_sint_sat/uint_sat.

The default fixed vector legalization is to unroll. The default
scalable vector legalization is to clamp in the FP domain. The
RVV vfcvt instructions have saturating behavior so we can use them
directly. The only difference is that RVV instruction turn nan into
the max value, but the _SAT intrinsics want 0.

I'm only supporting 1 step of narrowing for now. I think we can
support more steps by using VNCLIP to saturate and narrower.

The only case that needs 2 steps of widening is f16->i64 which we can
do as f16->f32->i64.

Reviewed By: frasercrmck

Differential Revision: https://reviews.llvm.org/D134400

21 months ago[RISCV] Formatting fixes to RISCV.td NFC
Craig Topper [Thu, 22 Sep 2022 02:04:48 +0000 (19:04 -0700)]
[RISCV] Formatting fixes to RISCV.td NFC

Improve indentation. Fix the worst of the 80 column violations.

21 months agoFix -Wunused-local-typedef warning in some build configurations
Dmitri Gribenko [Thu, 22 Sep 2022 15:09:58 +0000 (17:09 +0200)]
Fix -Wunused-local-typedef warning in some build configurations

21 months ago[GlobalISel] Enforce G_ASSERT_ALIGN to have a valid alignment > 0.
Amara Emerson [Thu, 22 Sep 2022 14:41:27 +0000 (15:41 +0100)]
[GlobalISel] Enforce G_ASSERT_ALIGN to have a valid alignment > 0.

21 months ago[BasicAA] Move experimental.guard modelling to getModRefBehavior()
Nikita Popov [Wed, 14 Sep 2022 15:10:39 +0000 (17:10 +0200)]
[BasicAA] Move experimental.guard modelling to getModRefBehavior()

While we can't express this with attributes yet, we can model
these intrinsics as readonly + writing inaccessible memory (for
the control dependence) in FMRB. This way we don't need to
special-case them in getModRefInfo(), it falls out of the usual
logic.

21 months ago[InstCombine] Remove buggy zext of icmp eq with pow2 fold (PR57899)
Nikita Popov [Thu, 22 Sep 2022 14:35:07 +0000 (16:35 +0200)]
[InstCombine] Remove buggy zext of icmp eq with pow2 fold (PR57899)

For the case where the constant is a power of two rather than zero,
the fold is incorrect, because it fails to check that the bit set
in the LHS matches the bit in the RHS.

Rather than fixing this, remove the power of two handling entirely,
as a different fold will already canonicalize such comparisons to
use a zero constant.

Fixes https://github.com/llvm/llvm-project/issues/57899.

21 months ago[Flang] Finalize IO operations after calling EndIO
Peter Steinfeld [Wed, 21 Sep 2022 20:08:02 +0000 (13:08 -0700)]
[Flang] Finalize IO operations after calling EndIO

The process of passing arguments to IO calls can cause allocations that
get referenced during EndIO calls.  Calling "Finalize" causes these
allocations to be deallocated.  This means that references to them in
the code in EndIO will be invalid.  The fix is to delay the call to
"finalize" until after the call to EndIO.

This particularly causes problems with the IO items are strings that are
produced by calls to functions.

Differential Revision: https://reviews.llvm.org/D134383

21 months ago[CostModel][X86] Add gep.ll CostKind test coverage
Simon Pilgrim [Thu, 22 Sep 2022 13:59:02 +0000 (14:59 +0100)]
[CostModel][X86] Add gep.ll CostKind test coverage

21 months ago[CostModel][X86] Regenerate gep.ll test checks
Simon Pilgrim [Thu, 22 Sep 2022 13:51:23 +0000 (14:51 +0100)]
[CostModel][X86] Regenerate gep.ll test checks

21 months ago[InstCombine] Add test for PR57899 (NFC)
Nikita Popov [Thu, 22 Sep 2022 13:59:43 +0000 (15:59 +0200)]
[InstCombine] Add test for PR57899 (NFC)

21 months ago[clang][docs] Fix supported element types for __builtin_reduce_(add|mul)
Joe Loser [Tue, 20 Sep 2022 21:15:26 +0000 (15:15 -0600)]
[clang][docs] Fix supported element types for __builtin_reduce_(add|mul)

The docs mention that `__builtin_reduce_add` and `__builtin_reduce_mul` support
both integer and floating point element types, but only integer element types
are actually supported. See https://github.com/llvm/llvm-project/issues/57847,
and specifically,
https://github.com/llvm/llvm-project/blob/00874c48ea4d291908517afaab50d1dcbfb016c3/clang/lib/Sema/SemaChecking.cpp#L2631 for the fact that floating point element types are not supported yet.

Fix the docs to only mention support for integer element types.

21 months agoMachineVerifier: Verify REG_SEQUENCE
Matt Arsenault [Wed, 21 Sep 2022 14:42:03 +0000 (10:42 -0400)]
MachineVerifier: Verify REG_SEQUENCE

Somehow there was no verification of this, other than an ad-hoc
assertion in TwoAddressInstructions.

21 months ago[InstCombine] Use simplifyWithOpReplaced() for non-bool selects
Nikita Popov [Thu, 22 Sep 2022 09:04:22 +0000 (11:04 +0200)]
[InstCombine] Use simplifyWithOpReplaced() for non-bool selects

Perform the simplifyWithOpReplaced() fold even for non-bool
selects. This subsumes a number of recently added folds for
zext/sext of the condition.

We still need to manually handle variations with both sext/zext
and not, because simplifyWithOpReplaced() only performs one
level of replacements.

21 months agoRe-apply "Deferred Concept Instantiation Implementation"
Erich Keane [Fri, 19 Aug 2022 20:57:45 +0000 (13:57 -0700)]
Re-apply "Deferred Concept Instantiation Implementation"

This reverts commit 95d94a67755620c0a2871ac6f056ca8e9731d5e9.

This implements the deferred concepts instantiation, which should allow
the libstdc++ ranges to properly compile, and for the CRTP to work for
constrained functions.

Since the last attempt, this has fixed the issues from @wlei and
@mordante.

Differential Revision: https://reviews.llvm.org/D126907

21 months ago[AST] Better recovery on an expression refers to an invalid decl.
Haojian Wu [Thu, 22 Sep 2022 09:50:17 +0000 (11:50 +0200)]
[AST] Better recovery on an expression refers to an invalid decl.

Prior to the patch, we didn't build a DeclRefExpr if the Decl being
referred to is invalid, because many clang downstream AST consumers
assume it, violating it will cause many diagnostic regressions.

With this patch, we build a DeclRefExpr enven for an invalid decl (when the
AcceptInvalidDecl is true), and wrap it with a dependent-type
RecoveryExpr (to prevent follow-up semantic analysis, and diagnostic
regressions).

This is a revised version of https://reviews.llvm.org/D76831

Reviewed By: sammccall

Differential Revision: https://reviews.llvm.org/D121599

21 months ago[LLD] [ELF] Fix building with LLVM_LINK_LLVM_DYLIB since zstd was taken into use
Martin Storsjö [Thu, 22 Sep 2022 11:08:45 +0000 (11:08 +0000)]
[LLD] [ELF] Fix building with LLVM_LINK_LLVM_DYLIB since zstd was taken into use

This fixes a regression since fa74144c64dff6b145b0b3fa9397f913ddaa87bf;
even if we're linking to the dylib (which handles all the dependencies
in LLVMSupport), we're now also directly referencing zstd from
lld/ELF, and thus need to explicitly express our dependency on it.

21 months ago[clang] Rework IsTailPaddedMemberArray into isFlexibleArrayMemberExpr
serge-sans-paille [Thu, 1 Sep 2022 13:15:29 +0000 (15:15 +0200)]
[clang] Rework IsTailPaddedMemberArray into isFlexibleArrayMemberExpr

This fixes a bunch of FIXME within IsTailPaddedMemberArray related code.

As a side effect, this now also triggers a warning when trying to access a
"struct hack" member with an index above address space index range.

Differential Revision: https://reviews.llvm.org/D133108

21 months ago[MemorySSA] Reset location size if IsGuaranteedLoopInvariant after phi tranlation
luxufan [Mon, 19 Sep 2022 03:29:16 +0000 (03:29 +0000)]
[MemorySSA] Reset location size if IsGuaranteedLoopInvariant after phi tranlation

We set the Location size to beforeOrAfter if the Location value is not
guaranteed loop invariant. But in some cases, we need to reset the
location size if the location size is precise after phi tranlation of
location value. This will improve MemorySSA analysis results.

Differential Revision: https://reviews.llvm.org/D134161

21 months ago[mlir] Fix a cast that should be a dyn_cast.
Johannes Reifferscheid [Thu, 22 Sep 2022 11:01:28 +0000 (13:01 +0200)]
[mlir] Fix a cast that should be a dyn_cast.

This fixes a crash for certain IR, see the new test case for an
example.

Reviewed By: ftynse

Differential Revision: https://reviews.llvm.org/D134424

21 months agoAArch64: add support for newer Apple CPUs
Tim Northover [Tue, 20 Sep 2022 14:03:26 +0000 (15:03 +0100)]
AArch64: add support for newer Apple CPUs

They're roughly ARMv8.6. This works in the .td file, but in
AArch64TargetParser.def, marking them v8.6 brings in support for the SM4
cryptographic hash and we don't actually have that. So TargetParser side
they're marked as v8.5, with the extra features (BF16 and I8MM added manually).

Finally, A16 supports the HCX extension in addition to v8.6. This has no
TargetParser implications.

21 months ago[CostModel][X86] Add partial CostKinds handling for funnelshifts/rotates
Simon Pilgrim [Thu, 22 Sep 2022 10:24:01 +0000 (11:24 +0100)]
[CostModel][X86] Add partial CostKinds handling for funnelshifts/rotates

This mainly just adds costs for the targets where we have actual funnelshift/rotate instructions (VBMI2/XOP etc.) - the cases where we expand still need addressing, although for many the default shift+or expansion, especially for uniform cases, isn't that bad.

This was achieved with the 'cost-tables vs llvm-mca' script D103695

21 months agoRe-land "[llvm-exegesis] Support analyzing results from a different target."
Clement Courbet [Thu, 22 Sep 2022 09:27:44 +0000 (11:27 +0200)]
Re-land "[llvm-exegesis] Support analyzing results from a different target."

With Mips fixes.

This reverts commit 7daf60e3440b22b79084bb325d823aa3ad8df0f3.

21 months ago[CostModel][X86] Add CostKinds handling for smax/smin/umax/umin instructions
Simon Pilgrim [Thu, 22 Sep 2022 09:19:16 +0000 (10:19 +0100)]
[CostModel][X86] Add CostKinds handling for smax/smin/umax/umin instructions

This was achieved with the 'cost-tables vs llvm-mca' script D103695

21 months agoRevert "[llvm-exegesis] Support analyzing results from a different target."
Clement Courbet [Thu, 22 Sep 2022 09:18:12 +0000 (11:18 +0200)]
Revert "[llvm-exegesis] Support analyzing results from a different target."

Breaks MIPS compile.

This reverts commit cc61c822e05c51e494c50d1e72f963750116ef08.

21 months ago[AA] Model operand bundles more precisely
Nikita Popov [Thu, 1 Sep 2022 14:51:00 +0000 (16:51 +0200)]
[AA] Model operand bundles more precisely

Based on D130896, we can model operand bundles more precisely. In
addition to the baseline ModRefBehavior, a reading/clobbering operand
bundle may also read/write all locations. For example, a memcpy with
deopt bundle can read any memory, but only write argument memory.

This means that getModRefInfo() for memcpy with a pointer that does
not alias the arguments results in Ref, rather than ModRef, without
the need to implement any special handling.

Differential Revision: https://reviews.llvm.org/D130980

21 months ago[llvm-exegesis] Support analyzing results from a different target.
Clement Courbet [Mon, 12 Sep 2022 08:13:14 +0000 (10:13 +0200)]
[llvm-exegesis] Support analyzing results from a different target.

We were using the native triple to parse the benchmarks. Use the triple
from the benchmarks file.

Right now this still only allows analyzing files produced by the current
target until D133605 is in.

This also makes the `Analysis` class much less ad-hoc.

Differential Revision: https://reviews.llvm.org/D133697

21 months ago[pseudo] NFC, Remove an extral blank line.
Haojian Wu [Thu, 22 Sep 2022 09:07:25 +0000 (11:07 +0200)]
[pseudo] NFC, Remove an extral blank line.

21 months ago[InstSimplify] Support vectors in simplifyWithOpReplaced()
Nikita Popov [Wed, 21 Sep 2022 09:57:49 +0000 (11:57 +0200)]
[InstSimplify] Support vectors in simplifyWithOpReplaced()

We can handle vectors inside simplifyWithOpReplaced(), as long as
cross-lane operations are excluded. The equality can hold (or not
hold) for each vector lane independently, so we shouldn't use the
replacement value from other lanes.

I believe the only operations relevant here are shufflevector (where
all previous bugs were seen) and calls (which might use shuffle-like
intrinsics and would require more careful classification).

Differential Revision: https://reviews.llvm.org/D134348

21 months ago[BOLT] Fix (part of) dylib compatibility
serge-sans-paille [Tue, 20 Sep 2022 12:45:11 +0000 (14:45 +0200)]
[BOLT] Fix (part of) dylib compatibility

Non-LLVM components should not be listed as part of LLVM_LINK_COMPONENTS.

Differential Revision: https://reviews.llvm.org/D134278

21 months agoRevert "[bolt] Fix (part of) dylib compatibility"
serge-sans-paille [Thu, 22 Sep 2022 08:41:21 +0000 (10:41 +0200)]
Revert "[bolt] Fix (part of) dylib compatibility"

This reverts commit 34ad83d883cc4505412a7c3e1e3da74e5408aa82.

21 months ago[bolt] Fix (part of) dylib compatibility
serge-sans-paille [Tue, 20 Sep 2022 12:45:11 +0000 (14:45 +0200)]
[bolt] Fix (part of) dylib compatibility

Non-LLVM component should not be listed as part of LLVM_LINK_COMPONENTS

Differential Revision: https://reviews.llvm.org/D134278

21 months ago[MLIR] Revert default NVIDIA GPU version
Ivan Radanov Ivanov [Thu, 22 Sep 2022 08:11:41 +0000 (10:11 +0200)]
[MLIR] Revert default NVIDIA GPU version

Due to integration tests failing revert mlir::SerializeToCubinPass defaults to old ones (changed in https://reviews.llvm.org/D134153)

Reviewed By: akuegel

Differential Revision: https://reviews.llvm.org/D134414

21 months ago[SPIRV] fix build with clang and use PoisonValue instead of UndefValue
Ilia Diachkov [Tue, 20 Sep 2022 13:40:57 +0000 (16:40 +0300)]
[SPIRV] fix build with clang and use PoisonValue instead of UndefValue

The patch fixes the SPIRV backend build using clang. It also replaces
UndefValue with PoisonValue in SPIRVRegularizer.cpp.

Fixes: #57773

Differential Revision: https://reviews.llvm.org/D134071

21 months ago[llvm-reduce] Add test with debug nodes not being removed.
Florian Hahn [Thu, 22 Sep 2022 07:49:56 +0000 (08:49 +0100)]
[llvm-reduce] Add test with debug nodes not being removed.

Add a test where llvm-reduce at the moment cannot drop unneeded debug
metadata nodes referenced by DICompileUnit and DISubpgoram.

21 months ago[gn build] Port 32a8260ccc8b
LLVM GN Syncbot [Thu, 22 Sep 2022 07:29:04 +0000 (07:29 +0000)]
[gn build] Port 32a8260ccc8b

21 months ago[clang] Fix interaction between asm labels and inline builtins
serge-sans-paille [Wed, 21 Sep 2022 14:08:45 +0000 (16:08 +0200)]
[clang] Fix interaction between asm labels and inline builtins

One must pick the same name as the one referenced in CodeGenFunction when
generating .inline version of an inline builtin, otherwise they are not
correctly replaced.

Differential Revision: https://reviews.llvm.org/D134362

21 months ago-dot-machine-cfg for printing MachineFunction to a dot file
Christudasan Devadasan [Thu, 22 Sep 2022 05:19:46 +0000 (10:49 +0530)]
-dot-machine-cfg for printing MachineFunction to a dot file

This pass allows a user to dump a MIR function to a dot file
and view it as a graph. It is targeted to provide a similar
functionality as -dot-cfg pass on LLVM-IR. As of now the pass
also support below flags:
-dot-mcfg-only [optional][won't print instructions in the
graph just block name]
-mcfg-dot-filename-prefix [optional][prefix to add to output dot file]
-mcfg-func-name [optional] [specify function name or it's
substring, handy if mir file contains multiple functions and
you need to see graph of just one]

More flags and details can be introduced as per the requirements
in future. This pass is inspired from -dot-cfg IR pass and APIs
are written in almost identical format.

Patch by Yashwant Singh <Yashwant.Singh@amd.com> (yassingh)

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D133709

21 months ago[InstCombine] Precommit tests for shl_trunc_icmp. nfc
Chenbing Zheng [Thu, 22 Sep 2022 07:08:38 +0000 (15:08 +0800)]
[InstCombine] Precommit tests for shl_trunc_icmp. nfc

21 months ago[NFC] Remove unused set construction from DILocation::getMergedLocation
Juan Manuel MARTINEZ CAAMAÑO [Wed, 21 Sep 2022 13:31:40 +0000 (13:31 +0000)]
[NFC] Remove unused set construction from DILocation::getMergedLocation

Reviewed By: dblaikie

Differential Revision: https://reviews.llvm.org/D134357

21 months ago[RelocationResolver] Add R_RISCV_SET{16,32}
LiaoChunyu [Thu, 22 Sep 2022 03:53:24 +0000 (11:53 +0800)]
[RelocationResolver] Add R_RISCV_SET{16,32}

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D134408

21 months ago[RISCV] Replace RISCVISD::GREV/GORC/SHFL/UNSHFL with BREV8/ORC_B/ZIP/UNZIP.
Craig Topper [Thu, 22 Sep 2022 04:54:03 +0000 (21:54 -0700)]
[RISCV] Replace RISCVISD::GREV/GORC/SHFL/UNSHFL with BREV8/ORC_B/ZIP/UNZIP.

With Zbp removed, we no longer need the generalized forms.

The computeKnownBitsForTargetNode code brev8/orc.b is still based
on the general form with the shift amount forced to 7.

21 months ago[RISCV] Remove support for the unratified Zbp extension.
Craig Topper [Thu, 22 Sep 2022 04:04:10 +0000 (21:04 -0700)]
[RISCV] Remove support for the unratified Zbp extension.

This extension does not appear to be on its way to ratification.

Still need some follow up to simplify the RISCVISD nodes.

21 months ago[LoopInterchange][PR57148] Ensure the correct form of IR after transformation
Congzhe Cao [Thu, 22 Sep 2022 04:00:52 +0000 (00:00 -0400)]
[LoopInterchange][PR57148] Ensure the correct form of IR after transformation

This is a bugfix patch that resolves the following two bugs in loop interchange:

1. PR57148 which is an assertion error due to of loss of LCSSA form after interchange,
   as referred to test1() in pr57148.ll.
2. Use before def for the outermost loop induction variables after interchange,
   as referred to test2() in pr57148.ll.

The fix in this patch is that:

1. In cases where the LCSSA form is not maintained after interchange, we update the IR
   to the LCSSA form again.
2. We split the phi nodes in the inner loop header into a separate basic block to avoid
   the situation where use of the outer indvar appears before its def after interchange.
   Previously we already did this for innermost loops, now we do it for non-innermost
   loops (e.g., middle loops) as well.

Reviewed By: bmahjour, Meinersbur, #loopoptwg

Differential Revision: https://reviews.llvm.org/D132055

21 months ago[WebAssembly] Improve codegen for loading scalars from memory to v128
Fanchen Kong [Thu, 22 Sep 2022 04:05:44 +0000 (21:05 -0700)]
[WebAssembly] Improve codegen for loading scalars from memory to v128

Use load32_zero instead of load32_splat to load the low 32 bits from memory to
v128. Test cases are added to cover this change.

Reviewed By: tlively

Differential Revision: https://reviews.llvm.org/D134257

21 months ago[LoopPassManager] Ensure to construct loop nests with the outermost loop
Congzhe Cao [Thu, 22 Sep 2022 03:35:10 +0000 (23:35 -0400)]
[LoopPassManager] Ensure to construct loop nests with the outermost loop

This patch is to resolve the bug reported and discussed in
https://reviews.llvm.org/D124926#3718761 and https://reviews.llvm.org/D124926#3719876.

The problem is that loop interchange is a loopnest pass under the new pass manager,
but the loop nest may not be constructed correctly by the loop pass manager after
running loop interchange and before running the next pass, which might cause problems
when it continues running the next pass.

The reason that the loop nest is constructed incorrectly is that the outermost
loop might have changed after interchange, and what was the original outermost
loop is not the current outermost loop anymore. Constructing the loop nest based
on the original outermost loop would generate an invalid loop nest.

The fix in this patch is that, in the loop pass manager before running each loopnest
pass, we re-cosntruct the loop nest based on the current outermost loop, if LPMUpdater
notifies the loop pass manager that the previous loop nest has been invalidated by passes
like loop interchange.

Reviewed By: aeubanks

Differential Revision: https://reviews.llvm.org/D132199

21 months ago[RelocationResolver] Add R_RISCV_SET8
LiaoChunyu [Wed, 21 Sep 2022 13:46:11 +0000 (21:46 +0800)]
[RelocationResolver] Add R_RISCV_SET8

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D134164

21 months ago[C++] [Coroutines] Prefer aligned (de)allocation for coroutines -
Chuanqi Xu [Thu, 22 Sep 2022 02:59:39 +0000 (10:59 +0800)]
[C++] [Coroutines] Prefer aligned (de)allocation for coroutines -
implement the option2 of P2014R0

This implements the option2 of
https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p2014r0.pdf.

This also fixes https://github.com/llvm/llvm-project/issues/56671.

Although wg21 didn't get consensus for the direction of the problem,
we're happy to have some implementation and user experience first. And
from issue56671, the option2 should be the pursued one.

Reviewed By: ychen

Differential Revision: https://reviews.llvm.org/D133341

21 months agoUpdate Unicode to 15.0
Corentin Jabot [Tue, 13 Sep 2022 18:09:56 +0000 (20:09 +0200)]
Update Unicode to 15.0

Unicode 15.0 adds 4,489 characters, for a total of 149,186 characters.
These additions include 2 new scripts along with 20 new emoji characters,
and 4,193 CJK ideographs.

This changes modify most existing tables including
 - XID_Start/XID_Continue in Clang
 - The character name database (used by \N{} in Clang)
 - The list of formattable/printable codepoints
 - The case folding algorithm (which we had not updated since Unicode 9)
 - The list of nonspacing/enclosing marks used by the column width
   computation algorithm. The rest of the column width algorithm
   is not updated.

Reviewed By: tahonermann

Differential Revision: https://reviews.llvm.org/D133807

21 months ago[RISCV] Remove some unused Predicates from tablegen. NFC
Craig Topper [Thu, 22 Sep 2022 01:17:22 +0000 (18:17 -0700)]
[RISCV] Remove some unused Predicates from tablegen. NFC

Specifically predicates for extensions that are subsets of other
extensions. These predicates should never be used. Should always
check the superset extension or the superset ORed with the sub extendsion.

21 months ago[mlir] Flip PDL to use Both accessors
River Riddle [Wed, 21 Sep 2022 20:40:47 +0000 (13:40 -0700)]
[mlir] Flip PDL to use Both accessors

This allows for incrementally updating the old API usages without
needing to update everything at once. PDL will be left on Both
for a little bit and then flipped to prefixed when all APIs have been
updated.

Differential Revision: https://reviews.llvm.org/D134387

21 months ago[mlir] Flip Async/GPU/OpenACC/OpenMP to use Both accessors
River Riddle [Wed, 21 Sep 2022 20:08:26 +0000 (13:08 -0700)]
[mlir] Flip Async/GPU/OpenACC/OpenMP to use Both accessors

This allows for incrementally updating the old API usages without
needing to update everything at once. These will be left on Both
for a little bit and then flipped to prefixed when all APIs have been
updated.

Differential Revision: https://reviews.llvm.org/D134386

21 months ago[analyzer] Support implicit parameter 'self' in path note
Jan Korous [Tue, 13 Sep 2022 21:45:45 +0000 (14:45 -0700)]
[analyzer] Support implicit parameter 'self' in path note

showBRParamDiagnostics assumed stores happen only via function parameters while that
can also happen via implicit parameters like 'self' or 'this'.
The regression test caused a failed assert in the original cast to ParmVarDecl.

Differential Revision: https://reviews.llvm.org/D133815

21 months ago[msan] Instrument vtest instrinsics
Vitaly Buka [Sun, 18 Sep 2022 06:44:23 +0000 (23:44 -0700)]
[msan] Instrument vtest instrinsics

Instrumentation just ORs shadow of inputs.
I assume some result shadow bits can be reset if we go into specifics of particular checks,
but as-is it is still an improvement against existing default strict instruction handler, when
every set bit of input shadow is  reported as an error.

Reviewed By: kda

Differential Revision: https://reviews.llvm.org/D134123

21 months ago[RISCV] Add a couple more fp_to_sint_sat/uint_sat tests. NFC
Craig Topper [Wed, 21 Sep 2022 23:44:02 +0000 (16:44 -0700)]
[RISCV] Add a couple more fp_to_sint_sat/uint_sat tests. NFC

21 months ago[RISCV] Add tests for fixed vector fp_to_sint_sat/fp_to_sint_sat. NFC
Craig Topper [Wed, 21 Sep 2022 23:16:37 +0000 (16:16 -0700)]
[RISCV] Add tests for fixed vector fp_to_sint_sat/fp_to_sint_sat. NFC

21 months ago[test] Fix CodeGen/arm-crc32.c to not write an output to CWD
Fangrui Song [Wed, 21 Sep 2022 23:13:25 +0000 (16:13 -0700)]
[test] Fix CodeGen/arm-crc32.c to not write an output to CWD

CWD may be read-only in some testing environments.

21 months ago[flang] Write semantics test for atomic_xor
Katherine Rasmussen [Tue, 20 Sep 2022 18:48:57 +0000 (11:48 -0700)]
[flang] Write semantics test for atomic_xor

Write a semantics test for the atomic intrinsic subroutine,
atomic_xor.

Reviewed By: rouson

Differential Revision: https://reviews.llvm.org/D134298

21 months ago[NFC] Make dxil namespace consistent
Chris Bieneman [Wed, 21 Sep 2022 22:34:39 +0000 (17:34 -0500)]
[NFC] Make dxil namespace consistent

We have namespaces `DXIL` and `dxil`, which is just confusing. This
renames `DXIL` -> `dxil` making everything consistent.

While the LLVM coding standards don't have a clear direction here, I
chose lower case because by my current unscientific count there are
more places where we had the lowercase namespace than the uppercase.

21 months agoRemove the unused/undefined `_cmd` parameter in `objc_direct` methods.
Michael Wyman [Wed, 14 Sep 2022 21:42:30 +0000 (14:42 -0700)]
Remove the unused/undefined `_cmd` parameter in `objc_direct` methods.

When `objc_direct` methods were implemented, the implicit `_cmd` parameter was left as an argument to the method implementation function, but was unset by callers; if the method body referenced the `_cmd` variable, a selector load would be emitted inside the body. However, this leaves an unused argument in the ABI, and is unnecessary.

This change removes the empty/unset argument, and if `_cmd` is referenced inside an `objc_direct` method it will emit local storage for the implicit variable. From the ABI perspective, `objc_direct` methods will have the implicit `self` parameter, immediately followed by whatever explicit arguments are defined on the method, rather than having one unset/undefined register in the middle.

Differential Revision: https://reviews.llvm.org/D131424

21 months ago[RISCV] Rename RISCVISD::SINT_TO_FP_VL/UINT_TO_FP_VL. NFC
Craig Topper [Wed, 21 Sep 2022 22:31:45 +0000 (15:31 -0700)]
[RISCV] Rename RISCVISD::SINT_TO_FP_VL/UINT_TO_FP_VL. NFC

Name them after the instructions VFCVT_RTZ_X(U)_F_VL to make it
clear that the ISD nodes don't have the poison semantics of
ISD::SINT_TO_FP/UINT_TO_FP.

I play to reuse this node for a FP_TO_SINT_SAT/FP_TO_UINT_SAT
patch and need the instruction semantics.

21 months ago[msan] Handle x86_avx_cmp_pd_256 and x86_avx_cmp_ps_256
Vitaly Buka [Sun, 18 Sep 2022 03:17:09 +0000 (20:17 -0700)]
[msan] Handle x86_avx_cmp_pd_256 and x86_avx_cmp_ps_256

Removed FIXME which looks irrelevant. The error message happens only without -mattr=+avx.
E.g.
GOOD: opt llvm/test/Instrumentation/MemorySanitizer/avx-intrinsics-x86.ll -passes=msan -o - | llc -O3 -o /dev/null -mattr=+avx
BAD: opt llvm/test/Instrumentation/MemorySanitizer/avx-intrinsics-x86.ll -passes=msan -o - | llc -O3 -o /dev/null

So nothing to fix here.

Reviewed By: RKSimon

Differential Revision: https://reviews.llvm.org/D134119

21 months ago[MLIR] Set default NVIDIA GPU version
Ivan Radanov Ivanov [Wed, 21 Sep 2022 22:10:10 +0000 (18:10 -0400)]
[MLIR] Set default NVIDIA GPU version

21 months ago[MLIR] Add a create function for mlir::SerializeToCubinPass
Ivan Radanov Ivanov [Wed, 21 Sep 2022 21:56:02 +0000 (17:56 -0400)]
[MLIR] Add a create function for mlir::SerializeToCubinPass

Differential Revision: https://reviews.llvm.org/D134153

21 months ago[Analysis] Make members of InlineCost const (NFC)
Kazu Hirata [Wed, 21 Sep 2022 21:47:23 +0000 (14:47 -0700)]
[Analysis] Make members of InlineCost const (NFC)

Once we create an instance of InlineCost, we don't change its
contents.

Differential Revision: https://reviews.llvm.org/D134388

21 months ago[TTI] Recognize fp constants in getOperandInfo
Philip Reames [Wed, 21 Sep 2022 21:33:04 +0000 (14:33 -0700)]
[TTI] Recognize fp constants in getOperandInfo

We were recognizing vectors of floats, but not scalars.  That's a tad odd.

21 months ago[test][msan] Precommit part of D134119
Vitaly Buka [Sun, 18 Sep 2022 03:17:09 +0000 (20:17 -0700)]
[test][msan] Precommit part of D134119

21 months ago[Hexagon] Fix -Wunused-variable in non-assertion builds after f6e7ad56040ee8531be3375...
Fangrui Song [Wed, 21 Sep 2022 21:14:45 +0000 (14:14 -0700)]
[Hexagon] Fix -Wunused-variable in non-assertion builds after f6e7ad56040ee8531be3375e04312d30700a58b2

21 months ago[libunwind] Handle G in personality string
Florian Mayer [Fri, 3 Jun 2022 21:33:08 +0000 (14:33 -0700)]
[libunwind] Handle G in personality string

Tested with the following program:

```
static volatile int* x = nullptr;

void throws()  __attribute__((noinline)) {
  if (getpid() == 0)
    return;
  throw "error";
}

void maybe_throws()  __attribute__((noinline)) {
  volatile int y = 1;
  x = &y;
  throws();
  y = 2;
}

int main(int argc, char** argv) {
  int y;
  try {
    maybe_throws();
  } catch (const char* e) {
    //printf("Caught\n");
  }
  y = *x;
  printf("%d\n", y); // should be MTE failure.
  return 0;
}
```

Built using `clang++ -c -O2 -target aarch64-linux -fexceptions -march=armv8-a+memtag -fsanitize=memtag-heap,memtag-stack`

Currently only Android implements runtime support for MTE stack tagging.

Without this change, we crash on `__cxa_get_globals` when trying to catch
the exception (because the stack frame __cxa_get_globals frame will fail due
to tags left behind on the stack). With this change, we crash on the `y = *x;`
as expected, because the stack frame has been untagged, but the pointer hasn't.

Reviewed By: #libunwind, compnerd, MaskRay

Differential Revision: https://reviews.llvm.org/D128998

21 months ago[Driver] Add --gcc-install-dir=
Fangrui Song [Wed, 21 Sep 2022 21:11:15 +0000 (14:11 -0700)]
[Driver] Add --gcc-install-dir=

This option specifies a GCC installation directory such as
/usr/lib/gcc/x86_64-linux-gnu/12, /usr/lib/gcc/x86_64-gentoo-linux-musl/11.2.0 .

It is intended to replace --gcc-toolchain=, which specifies a directory where
`lib/gcc{,-cross}` can be found. When --gcc-toolchain= is specified, the
selected `lib/gcc/$triple/$version` installation uses complex logic and the
largest GCC version is picked. There is no way to specify another version in the
presence of multiple GCC versions.

D25661 added gcc-config detection for Gentoo: `ScanGentooConfigs`.
The implementation may be simplified by using --gcc-install-dir=.

Reviewed By: mgorny

Differential Revision: https://reviews.llvm.org/D133329

21 months ago[GlobalISel] Fix std::max after D134380
Fangrui Song [Wed, 21 Sep 2022 21:09:04 +0000 (14:09 -0700)]
[GlobalISel] Fix std::max after D134380

21 months agoAMDGPU: Stop validating earlyclobber operands in assembler
Jay Foad [Tue, 20 Sep 2022 09:43:40 +0000 (10:43 +0100)]
AMDGPU: Stop validating earlyclobber operands in assembler

This validation was introduced in D34003 for v_qsad/v_mqsad instructions
but it applies to all instructions with earlyclobber operands, which now
includes v_mad_i64/v_mad_u64.

In all these cases I do not think there is documentation saying that the
destination must not overlap the sources. Rather there are *some* cases
where the instruction may not function correctly if there is an overlap,
and we are using earlyclobber as a conservative way of preventing
codegen from generating those cases.

I think it is unhelpful for the assembler to enforce the earlyclobber
restriction because it prevents assembling cases where the programmer
knows that in fact the overlap is safe.

See also: https://github.com/llvm/llvm-project/issues/57610

Differential Revision: https://reviews.llvm.org/D134272

21 months ago[GlobalISel] Fix known bits for G_ASSERT_ALIGN.
Amara Emerson [Wed, 21 Sep 2022 19:37:31 +0000 (20:37 +0100)]
[GlobalISel] Fix known bits for G_ASSERT_ALIGN.

I don't know what was going on originally with these tests. It seems reasonable
to have the immediate be the same byte alignment unit as the IR, in which case
we need to take the log2 in order to set the right number of low bits.

This fixes a miscompile in chromium.

Differential Revision: https://reviews.llvm.org/D134380

21 months ago[clang-format][NFC] Reformat the clang/unittests/Format directory
owenca [Wed, 21 Sep 2022 20:17:28 +0000 (13:17 -0700)]
[clang-format][NFC] Reformat the clang/unittests/Format directory

Also add a .clang-format file to clang/include/clang/Format and
clang/unittests/Format to keep the directories formatted.

21 months agoRevert "[llvm] Support forward-referenced globals with dso_local_equivalent"
Leonard Chan [Wed, 21 Sep 2022 20:13:31 +0000 (20:13 +0000)]
Revert "[llvm] Support forward-referenced globals with dso_local_equivalent"

This reverts commit 411020ad1c15a5f8cdbf674a2a8f845000f761e3.

One of the tests here fails on some upstream builders:
https://lab.llvm.org/buildbot#builders/16/builds/35314

21 months ago[Clang] Support case and default labels at end of compound statement
Evgeny Shulgin [Mon, 19 Sep 2022 19:26:03 +0000 (19:26 +0000)]
[Clang] Support case and default labels at end of compound statement

Direct continuation of https://reviews.llvm.org/D133887

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

Differential Revision: https://reviews.llvm.org/D134207

21 months agoAdd tests & update the C99 DR statuses for dr315 and dr316
Aaron Ballman [Wed, 21 Sep 2022 19:31:53 +0000 (15:31 -0400)]
Add tests & update the C99 DR statuses for dr315 and dr316

21 months ago[llvm] Support forward-referenced globals with dso_local_equivalent
Leonard Chan [Mon, 19 Sep 2022 23:22:14 +0000 (23:22 +0000)]
[llvm] Support forward-referenced globals with dso_local_equivalent

See https://github.com/llvm/llvm-project/issues/57815.

dso_local_equivalent would fail with an assertion on forward-referenced
globals. This is an issue that only comes up in textual IR, which is why
we've never seen this assertion with clang.

Differential Revision: https://reviews.llvm.org/D134234

21 months ago[HLSL] Allow SV_GroupIndex for lib profile.
Xiang Li [Tue, 20 Sep 2022 23:40:55 +0000 (16:40 -0700)]
[HLSL] Allow SV_GroupIndex for lib profile.

Lib profile could include all kind of entry functions.

Reviewed By: beanz

Differential Revision: https://reviews.llvm.org/D134326

21 months ago[InstCombine] try to fold mul by neg-power-of-2 to shl
Sanjay Patel [Wed, 21 Sep 2022 18:36:27 +0000 (14:36 -0400)]
[InstCombine] try to fold mul by neg-power-of-2 to shl

`(A * -2**C) + B --> B - (A << C)`

https://alive2.llvm.org/ce/z/A6BWkf

This inverts what Negator was doing before:
D134310 / 0f32a5dea0e9

Analysis and codegen are generally better without multiply,
so we should favor this form even if we trade add for sub
(because those are generally equivalent cost operations).

21 months ago[clang-format][NFC] Reformat clang/lib/Format using 6257832bf94f
owenca [Wed, 21 Sep 2022 00:00:05 +0000 (17:00 -0700)]
[clang-format][NFC] Reformat clang/lib/Format using 6257832bf94f

Fix braces and add .clang-format to keep the directory formatted.

Differential Revision: https://reviews.llvm.org/D134329

21 months agoChanging some strange code into an assert; NFC
Aaron Ballman [Wed, 21 Sep 2022 18:51:17 +0000 (14:51 -0400)]
Changing some strange code into an assert; NFC

This code was added in b65b1f322bd88513586a4539d2b5f18aeb698f3f, but it
was not noticed that the [[fallthrough]] behavior was very wrong. In C
mode, we would set the ParenExprType to CompoundLiteral and then
promptly overwrite that information by falling through.

After some investigation, I convinced myself that it is not possible to
hit this code path in C, only in C++. I've switched it to be an
assertion; I don't expect to hit it, but if we do hit it, that will at
least give us a code example we can use to reason about the intent of
the original code.