Joshua Batista [Tue, 20 Sep 2022 19:46:06 +0000 (14:46 -0500)]
[HLSL] add sqrt library function
This change exposes the sqrt library function for HLSL scalar types,
excluding long and long long doubles. Sqrt is supported for all scalar, vector,
and matrix types. This patch only adds a subset of scalar type support.
Long and long long double support is missing in this patch because that type
doesn't exist in HLSL.
The full documentation of the HLSL asin function is available here:
https://docs.microsoft.com/en-us/windows/win32/direct3dhlsl/dx-graphics-hlsl-sqrt
Reviewed By: beanz
Differential Revision: https://reviews.llvm.org/D132711
Zequan Wu [Tue, 20 Sep 2022 20:38:18 +0000 (13:38 -0700)]
[LLDB][NativePDB] Pass register_kind as reference (an oversight in 71d778f).
Fix #57799
Scott Linder [Tue, 20 Sep 2022 20:05:19 +0000 (20:05 +0000)]
[NFC][AMDGPU] Refactor AMDGPUDisassembler
Clean up ahead of a patch to fix bugs in the AMDGPUDisassembler.
Use lit.local.cfg substitutions and more idiomatic use of split-file to
simplify and extend existing kernel-descriptor disassembly tests.
Add a comment to AMDHSAKernelDescriptor.h, as at least one small set
towards keeping all kernel-descriptor sensitive code in sync.
Reviewed By: kzhuravl, arsenm
Differential Revision: https://reviews.llvm.org/D130105
Craig Topper [Tue, 20 Sep 2022 19:13:01 +0000 (12:13 -0700)]
[RISCV] Support -mno-implicit-float.
This can be used to disable vectorization and memcpy/memset
expansion for things like OS kernels. It also disables implicit
uses of scalar FP, but I don't know if we have any of those for
RISC-V.
NOTE: Without this patch you can still do -Xclang -no-implicit-float
Reviewed By: rui.zhang
Differential Revision: https://reviews.llvm.org/D134077
Anshil Gandhi [Tue, 20 Sep 2022 20:23:46 +0000 (14:23 -0600)]
[AMDGPU] Fix size of SOPK instructions to 4 bytes
Instructions in SOPK format may not have 32-bit
literal constants following the instruction.
Differential Revision: https://reviews.llvm.org/D133972
Markus Böck [Tue, 20 Sep 2022 20:12:10 +0000 (22:12 +0200)]
[InstCombine] Handle integer extension in `select` patterns using the condition as value
These patterns were previously only implemented for i1 type but can be extended for any integer type by also handling zext and sext operands.
Differential Revision: https://reviews.llvm.org/D134142
Sam McCall [Tue, 20 Sep 2022 20:17:02 +0000 (22:17 +0200)]
[clangd] Remove second tracer (which breaks threading contract)
rkayaith [Tue, 20 Sep 2022 02:14:45 +0000 (22:14 -0400)]
[mlir][standalone] Specify python path when configuring
Specifying the python path here ensures that the python binary used matches the
one used by the main MLIR tests. This is useful when cmake's automatic detection
has to be overridden.
Reviewed By: stellaraccident, bondhugula
Differential Revision: https://reviews.llvm.org/D134251
Frederik Gossen [Tue, 20 Sep 2022 19:11:12 +0000 (15:11 -0400)]
[MLIR] Update bazel build files after
242d558658cd5a480b02883e2982d7246342e0d0
Differential Revision: https://reviews.llvm.org/D134301
George Hu [Tue, 20 Sep 2022 17:56:16 +0000 (10:56 -0700)]
[LLDB]Initialize accept_socket with nullptr
Fix high impact issue of illegal access of memory.
Initialize accept_socket with nullptr.
Differential Revision: https://reviews.llvm.org/D134293
owenca [Mon, 19 Sep 2022 22:28:18 +0000 (15:28 -0700)]
[clang-format] Wrap inserted braces only if preceded by comments
Fixes #57805.
Differential Revision: https://reviews.llvm.org/D134233
Zain Jaffal [Tue, 20 Sep 2022 18:45:02 +0000 (19:45 +0100)]
[InstCombine] Matrix multiplication negation optimisation
If one of the operands in a matrix multiplication is negated we can optimise the equation by moving the negation to the smallest element of the operands or the result.
Reviewed By: spatel, fhahn
Differential Revision: https://reviews.llvm.org/D133300
Matt Arsenault [Mon, 19 Sep 2022 21:50:36 +0000 (17:50 -0400)]
AMDGPU: Fix expansion of 16-bit atomicrmw
Fixes issue 57830
Douglas Yung [Tue, 20 Sep 2022 18:23:47 +0000 (11:23 -0700)]
Add explicit tests for the PS4/PS5 C/C++ standards version defaults.
Alex Brachet [Tue, 20 Sep 2022 18:25:16 +0000 (18:25 +0000)]
[Driver] Make --execute-only the default for aarch64-fuchsia
Clang already generates code that doesn't use writeable data in executable
sections so the linker flag is all that is necessary.
-Wl,--no-execute-only can be used to turn this default off.
Differential Revision: https://reviews.llvm.org/D134289
Katherine Rasmussen [Mon, 19 Sep 2022 17:53:10 +0000 (10:53 -0700)]
[flang] Write semantics test for atomic_or
Write a semantics test for the atomic intrinsic subroutine,
atomic_or.
Reviewed By: rouson
Differential Revision: https://reviews.llvm.org/D134200
Fangrui Song [Tue, 20 Sep 2022 18:14:04 +0000 (11:14 -0700)]
[ELF] Name ctx->objectFiles[0] in the EF_RISCV_FLOAT_ABI mismatch error
Reviewed By: compnerd
Differential Revision: https://reviews.llvm.org/D134198
Peiming Liu [Tue, 20 Sep 2022 17:50:34 +0000 (17:50 +0000)]
[mlir][scf] Support simple symbolic expression without depending on AffineDialect to simply trivial loops.
Remove dependence of AffineDialect
Reviewed By: mehdi_amini
Differential Revision: https://reviews.llvm.org/D134291
Jeff Niu [Wed, 14 Sep 2022 00:19:19 +0000 (17:19 -0700)]
[mlir][ods] Allow empty array ref parameter
This patch "fixes" a longstanding issue where the assembly format for
ArrayRefParameter could not handle an empty list. This is because there
was no way to generically optionally parse the first element of the
array. The only solution was to write a (relatively simple) custom parser.
This patch implements "empty" ArrayRefParameters by using
inverted optional groups and an optional ArrayRefParameter.
Depends on D133816
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D133819
Jeff Niu [Tue, 13 Sep 2022 23:16:20 +0000 (16:16 -0700)]
[mlir][ods] Generalize default/optional parameters
This patch consolidates the notions of an optional parameter and a
default parameter. An optional parameter is a parameter equal to its
default value, which for a "purely optional" parameter is its "null"
value.
This allows the existing `comparator` and `defaultValue` fields to be
used enabled more complex "optional" parameters, such as empty arrays.
Depends on D133812
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D133816
Jeff Niu [Tue, 13 Sep 2022 22:08:39 +0000 (15:08 -0700)]
[mlir][ods] Format: allow anchors in the else elements
This patch changes optional groups to allow anchors in the 'else'
element group. When printing, the optional condition is inverted to
decide which group to print. This is useful for parsing concrete
optional elements that don't have a `parseOptional*` method or some
other way to test whether it's present.
Depends on D133805
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D133812
Jeff Niu [Tue, 13 Sep 2022 20:15:06 +0000 (13:15 -0700)]
[mlir][ods] Store the pointer to the anchor element (NFC)
Instead of its index. There is no benefit to storing the index instead
of the pointer.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D133805
Brett Wilson [Tue, 20 Sep 2022 17:54:28 +0000 (17:54 +0000)]
[clang-doc] Centralize TypeInfo creation.
Several different places in the code had similar computations for the parameters that were eventually passed to the TypeInfo constructor.
This centralizes that code in one function, and allows passing TypeInfo to the various other *Info structures that need it.
Remove some "auto" types and replace with the real type for getting declarations. This was making some duplicate checking difficult to see.
Reviewed By: paulkirth
Differential Revision: https://reviews.llvm.org/D134225
Jorge Gorbe Moya [Mon, 19 Sep 2022 23:35:24 +0000 (16:35 -0700)]
[NFCI] Clean up enum FormatCategoryItem.
- Merge pairs like `eFormatCategoryItemSummary` and
`eFormatCategoryItemRegexSummary` into a single value. See explanation
below.
- Rename `eFormatCategoryItemValue` to `eFormatCategoryItemFormat`. This
makes the enum match the names used elsewhere for formatter kinds
(format, summary, filter, synth).
- Delete unused values `eFormatCategoryItemValidator` and
`eFormatCategoryItemRegexValidator`.
This enum is only used to reuse some code in CommandObjectType.cpp. For
example, instead of having separate implementations for `type summary
delete`, `type format delete`, and so on, there's a single generic
implementation that takes an enum value, and then the specific commands
derive from it and set the right flags for the specific kind of
formatter.
Even though the enum distinguishes between regular and regex matches for
every kind of formatter, this distinction is never used: enum values are
always specified in pairs like
`eFormatCategoryItemSummary | eFormatCategoryItemRegexSummary`.
This causes some ugly code duplication in TypeCategory.cpp. In order to
handle every flag combination some code appears 8 times:
{format, summary, synth, filter} x {exact, regex}
Differential Revision: https://reviews.llvm.org/D134244
Anton Sidorenko [Tue, 20 Sep 2022 17:22:16 +0000 (10:22 -0700)]
[NFC][RISCV] Move calculations of SDNode policy operand idx to a separate function
Since there is no guaranteed correspondence of SDNode and MI operands, we need
getters simular to RISCVII::get*OpNum for SDNodes.
More uses of getVecPolicyOpIdx will be added in D130895.
Reviewed By: craig.topper, arcbbb
Differential Revision: https://reviews.llvm.org/D134179
Philip Reames [Tue, 20 Sep 2022 17:01:37 +0000 (10:01 -0700)]
[RISCV][MC] Add support for experimental Zawrs extension
This implements experimental support for the Zawrs extension as specified here: https://github.com/riscv/riscv-zawrs/releases/download/V1.0-rc3/Zawrs.pdf. Despite the 1.0 version name, this has not been ratified and there was a major change to proposed specification between rc2 and rc3. Once this is ratified, it'll move out of experimental status.
This change adds assembly support, but does not include C language or IR intrinsics. We can decide if we want them, and handle that in a separate patch.
Differential Revision: https://reviews.llvm.org/D133443
LLVM GN Syncbot [Tue, 20 Sep 2022 17:07:57 +0000 (17:07 +0000)]
[gn build] Port
00798e500644
Mingming Liu [Tue, 14 Jun 2022 21:04:49 +0000 (14:04 -0700)]
[AArch64] Define __ARM_FEATURE_RCPC
This patch implements the definition of __ARM_FEATURE_RCPC when clang
command specifies +rcpc.
Differential Revision: https://reviews.llvm.org/D127798
Mark de Wever [Mon, 5 Sep 2022 17:01:50 +0000 (19:01 +0200)]
[libc++] Applies P0602R4 retro-actively.
While testing a test failure of C++17 with Clang ToT it was noticed the
paper
P0602R4 variant and optional should propagate copy/move triviality
was not applied as a DR in libc++.
This was discovered while investigating the issue "caused by" D131479.
Reviewed By: #libc, ldionne
Differential Revision: https://reviews.llvm.org/D133326
Gulfem Savrun Yeniceri [Fri, 16 Sep 2022 18:19:06 +0000 (18:19 +0000)]
[InstrProfiling] Emit runtime hook only once
This patch fixes the issue about calling emitRuntimeHook() twice
when we need to unconditionally emit runtime hook as discussed in
https://reviews.llvm.org/rGd6aed77f0d19.
Differential Revision: https://reviews.llvm.org/D134254
Mark de Wever [Fri, 16 Sep 2022 18:47:01 +0000 (20:47 +0200)]
[NFC][libc++][format] Uses ranges algorithm.
Reviewed By: #libc, ldionne
Differential Revision: https://reviews.llvm.org/D134060
Mark de Wever [Thu, 5 May 2022 06:03:58 +0000 (08:03 +0200)]
[libc++][test] Adds format string helper.
Update the formatter day tests to the new style.
Other test will be done separately.
Reviewed By: #libc, ldionne
Differential Revision: https://reviews.llvm.org/D134031
Mark de Wever [Tue, 13 Sep 2022 18:10:26 +0000 (20:10 +0200)]
[libc++][chrono] Removes format include.
Switch to the new granular format_functions header. Since the chrono's
format dependency in C++20 hasn't been in a release it's save to remove
it.
Depends on D133665
Reviewed By: #libc, ldionne
Differential Revision: https://reviews.llvm.org/D133796
Mark de Wever [Sun, 11 Sep 2022 11:05:26 +0000 (13:05 +0200)]
[libc++][format] Granularizes the format header.
Moves the last pieces of code to its own header.
Reviewed By: ldionne, #libc
Differential Revision: https://reviews.llvm.org/D133665
Jay Foad [Tue, 20 Sep 2022 12:29:39 +0000 (13:29 +0100)]
[AMDGPU] Disable fp atomic to s_denorm_mode hazard for GFX11
This hazard only exists on GFX10.
Differential Revision: https://reviews.llvm.org/D134276
yronglin [Tue, 20 Sep 2022 16:35:18 +0000 (12:35 -0400)]
Fix __builtin_assume_aligned incorrect type descriptor and C++ object polymorphic address
Fix __builtin_assume_aligned incorrect type descriptor
example from @rsmith
struct A { int n; };
struct B { int n; };
struct C : A, B {};
void *f(C *c) {
// Incorrectly returns `c` rather than the address of the B base class.
return __builtin_assume_aligned((B*)c, 8);
}
Differential Revision: https://reviews.llvm.org/D133583
Kazu Hirata [Tue, 20 Sep 2022 16:09:38 +0000 (09:09 -0700)]
[IPO] Reorder parameters of InlineFunction (NFC)
With the recent addition of new parameter MergeAttributes (D134117),
callers need to specify several default parameters before getting to
specify the new parameter.
This patch reorders the parameters so that callers do not have to
specify as many default parameters.
Differential Revision: https://reviews.llvm.org/D134125
David Green [Tue, 20 Sep 2022 16:09:14 +0000 (17:09 +0100)]
[AArch64] Enable LSLFast for modern OoO cpus
This patch enables the LSLFast feature for Cortex-A76, Cortex-A77,
Cortex-A78, Cortex-A78C, Cortex-A710, Cortex-X1, Cortex-X2, Neoverse N1,
Neoverse N2, Neoverse V1 and the Neoverse 512TB pseudo-cpu, in-line with
the software optimization guides for those CPUs.
Differntial revision: https://reviews.llvm.org/D134273
Chris Bieneman [Tue, 20 Sep 2022 15:41:43 +0000 (10:41 -0500)]
[HLSL] Pass flags to cc1 based on language
Having the flags only pass through if you're using the dxc-driver means
that the clang driver doesn't work for HLSL, which is undesirable. This
change switches to instead passing flags based on the language mode
similar to how OpenCL does it. This allows the clang driver to be used
for HLSL source files as well.
Reviewed By: python3kgae
Differential Revision: https://reviews.llvm.org/D133958
Ben Dunbobbin [Tue, 20 Sep 2022 15:50:18 +0000 (16:50 +0100)]
Revert "[Utils] Refactor update_cc_test_checks.py to use shutil"
This reverts commit
2e6c50855b7d879ded3776ba87b3d960e2035b89.
This caused failures on windows bots:
- https://lab.llvm.org/buildbot/#/builders/216/builds/10030
- https://ci.chromium.org/ui/p/fuchsia/builders/toolchain.ci/clang-windows-x64/
b8802513693562827489/overview
Simon Pilgrim [Tue, 20 Sep 2022 15:44:26 +0000 (16:44 +0100)]
[InstCombine] Fold ult(add(x,-1),c) -> ule(x,c) iff x != 0 (PR57635)
Alive2: https://alive2.llvm.org/ce/z/sZ6wwS
As detailed on Issue #57635 and #37628 - for unsigned comparisons, we can compare prior to a decrement iff the value is known never to be zero.
Differential Revision: https://reviews.llvm.org/D134172
Chris Bieneman [Mon, 19 Sep 2022 15:22:33 +0000 (10:22 -0500)]
[CMake] [NFC] Add clang headers to IDE projects
This just adds the clang headers into a source group so that they get
collected and added into generated IDE projects.
Amara Emerson [Mon, 19 Sep 2022 23:21:55 +0000 (00:21 +0100)]
[GlobalISel][Legalizer] Fix lowerSelect() not sign-extending the mask value.
I'm not sure why the SEXT_INREG was gated on a bitwidth check of the mask
vs element size.
This fixes a miscompile in chromium's skia library.
Differential Revision: https://reviews.llvm.org/D134236
Jakub Kuderski [Tue, 20 Sep 2022 15:37:26 +0000 (11:37 -0400)]
[mlir][arith] Add comparison-based integration tests
Introduces a simple framework for runtime tests of the wide integer emulation.
In these tests, we are only interested in checking that both wide and narrow calculation
produce the same results, and do not check for exact results. This allows us to cover
more of the input space, as we do not have to hardcode each of the expected outputs.
Introduce common helper functions to check the results, print a message on mismatch,
and sample the input space.
Implement runtime comparrison tests for `arith.muli` and `arith.shrui`.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D134184
Jakub Kuderski [Tue, 20 Sep 2022 15:27:48 +0000 (11:27 -0400)]
[mlir][arith] Add integration test for shrui emulation
The new test cases focus on known edge cases in the current implementation.
Specifically, we check for low (0, 1), mid (7, 8, 9) and high (15) shift amounts with i16 operands.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D134182
Jakub Kuderski [Tue, 20 Sep 2022 15:03:37 +0000 (11:03 -0400)]
[mlir][arith] Add test pass for wide integer emulation
The new test pass allows for running wide integer emulation conversion
within specified functions only.
I intend to use it in integration tests in a way that allows me print both
original and emulated results in the same format, or even compare both results
at runtime and print on mismatch only.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D134120
Louis Dionne [Tue, 23 Aug 2022 15:22:55 +0000 (11:22 -0400)]
[llvm] Remove libcxx, libcxxabi and libunwind from supported LLVM_ENABLE_PROJECTS
This is a breaking change. If you were passing one of those three runtimes
in LLVM_ENABLE_PROJECTS, you need to start passing them in LLVM_ENABLE_RUNTIMES
instead. The runtimes in LLVM_ENABLE_RUNTIMES will start being built using
the "bootstrapping build" instead, which means that they will be built
using the just-built Clang. This is usually what you wanted anyway.
If you were using LLVM_ENABLE_PROJECTS=all with the explicit goal of
building these three runtimes, you can now use LLVM_ENABLE_RUNTIMES=all
and these runtimes will be built using the bootstrapping build.
Differential Revision: https://reviews.llvm.org/D132480
Eric Li [Tue, 20 Sep 2022 15:10:16 +0000 (11:10 -0400)]
[Support][NFC] Clarify function comment
Follow-up to
86118ec2 that addresses the comments in D134072, which
were accidentally left off of the commit.
Eric Li [Fri, 16 Sep 2022 20:07:26 +0000 (16:07 -0400)]
[Support] Provide access to the full mapping in llvm::Annotations
Providing access to the mapping of annotations allows test helpers to
be expressive by using the annotations as expectations. For example, a
matcher could verify that all annotated points were matched by a
matcher, or that an refactoring surgically modifies specific ranges.
Differential Revision: https://reviews.llvm.org/D134072
Joe Nash [Wed, 13 Jul 2022 18:14:48 +0000 (14:14 -0400)]
[AMDGPU][GFX11] Use VGPR_32_Lo128 for VOP1,2,C
Due to the encoding changes in GFX11, we had a hack in place that
disables the use of VGPRs above 128. This patch removes the need for
that hack.
We introduce a new register class VGPR_32_Lo128 which is used for 16-bit
operands of VOP1, VOP2, and VOPC instructions. This register class only has the
low 128 VGPRs, but is otherwise identical to VGPR_32. Therefore, 16-bit VOP1,
VOP2, and VOPC instructions are correctly limited to use the first 128
VGPRs, while the other instructions can freely use all 256.
We introduce new pseduo-instructions used on GFX11 which have the suffix
t16 (True 16) to use the VGPR_32_Lo128 register class.
Reviewed By: foad, rampitec, #amdgpu
Differential Revision: https://reviews.llvm.org/D133723
Jeff Bailey [Tue, 20 Sep 2022 14:25:16 +0000 (14:25 +0000)]
[libc] Fix TWS issues in .td files
Reviewed By: sivachandra
Differential Revision: https://reviews.llvm.org/D134256
Jeff Bailey [Tue, 20 Sep 2022 14:24:21 +0000 (14:24 +0000)]
[libc] Remove unneeded extra include
Reviewed By: sivachandra
Differential Revision: https://reviews.llvm.org/D134255
Florian Hahn [Tue, 20 Sep 2022 14:20:27 +0000 (15:20 +0100)]
[LV] Remove unused widenCallInstruction declaration (NFC).
The definition and uses have been removed a while ago. Clean up the
unused declaration.
Vy Nguyen [Fri, 16 Sep 2022 17:38:20 +0000 (13:38 -0400)]
[lld-macho] Support -dyld_env
This arg is undocumented but from looking at the code + experiment, it's used to add additional DYLD_ENVIRONMENT load commands to the output.
Differential Revision: https://reviews.llvm.org/D134058
Matt Arsenault [Tue, 20 Sep 2022 13:26:32 +0000 (09:26 -0400)]
Analysis: Remove redundant assertion
This assert guards the same assertion inside getTypeStoreSizeInBits
Matt Arsenault [Tue, 20 Sep 2022 12:53:25 +0000 (08:53 -0400)]
Analysis: Pass AssumptionCache through isKnownNonZero
Pass this through now that isDereferenceableAndAlignedPointer has
access to this.
Simon Pilgrim [Tue, 20 Sep 2022 13:23:59 +0000 (14:23 +0100)]
Fix MSVC "result of 32-bit shift implicitly converted to 64 bits" warning. NFCI.
Markus Böck [Tue, 20 Sep 2022 12:37:05 +0000 (14:37 +0200)]
[InstCombine] Precommit test for D134142
Simon Pilgrim [Tue, 20 Sep 2022 12:36:09 +0000 (13:36 +0100)]
[InstCombine] Add test coverage for D134172 / Issue #57635
Caroline Concatto [Tue, 30 Aug 2022 10:17:08 +0000 (11:17 +0100)]
[LLVM][AArch64] Replace aarch64.sve.ld by aarch64.sve.ldN.sret
This patch removes the intrinsic aarch64.sve.ldN from tablegen in favour of
using arch64.sve.ldN.sret.
Depends on: D133023
Differential Revision: https://reviews.llvm.org/D133025
gonglingqin [Tue, 20 Sep 2022 11:05:08 +0000 (19:05 +0800)]
[LoongArch] Add codegen support for fmaxnum_ieee and fminnum_ieee
Thanks for @xry111's previous bug fixes.
See https://github.com/loongson/llvm-project/pull/1 for more details.
Differential Revision: https://reviews.llvm.org/D133478
Djordje Todorovic [Tue, 13 Sep 2022 13:13:58 +0000 (15:13 +0200)]
Recommit "[AggressiveInstCombine] Lower Table Based CTTZ
The bug reported on the [0] has been fixed.
The issue was we have not checked if the global variables that
represent cttz tables was constant.
There is a new negative test added in negative-lower-table-based-cttz.ll
that represents this.
[0] https://reviews.llvm.org/rGdf868edee561eb973edd85ec9df41c67aa0bff6b
Nikita Popov [Tue, 20 Sep 2022 11:06:35 +0000 (13:06 +0200)]
[LICM] Add test for PR57780 (NFC)
Blue Gaston [Fri, 16 Sep 2022 22:49:44 +0000 (15:49 -0700)]
[Sanitizer] Bump macOS deployment target for sanitizer unit test binary to support C++17 requirements.
This patch fixes a test failure on Apple caused by changing standard to c++17.
sanitizer_allocator_test.cpp requires language features introducied in 10.13 for c++17.
After initial investigation, it was not clear how to add this flag to a single file:
https://reviews.llvm.org/D133878
Becuase of this, we have upped the min version of this test suite to 10.13, the min version necessary to support necessary language features.
We felt this was a better option than upping the min version of the product to support a single test.
We are raising deployment target for a single test suite, rather than the product.
Differential Revision: https://reviews.llvm.org/D134091
rdar://
98737270
Tim Northover [Fri, 9 Sep 2022 09:46:23 +0000 (10:46 +0100)]
AAArch64: disable asynchronous unwind by default for MachO.
AArch64 MachO has a compact unwind format where most functions' unwind info can
be represented in just 4 bytes. But this cannot represent any asynchronous CFI
function, so it's essentially disabled when that's used. This is a large
code-size hit that we'd rather not take unless explicitly requested.
Tim Northover [Fri, 9 Sep 2022 09:01:02 +0000 (10:01 +0100)]
Refactor unwind table driver interface to expose default level. NFC.
Martin Storsjö [Tue, 20 Sep 2022 09:38:37 +0000 (12:38 +0300)]
[LLD] [test] Add a missing REQUIRES: x86
This was missed in
365d0a5cd867cdf414b70c9f4fd5122146287b01.
Simon Pilgrim [Tue, 20 Sep 2022 09:33:29 +0000 (10:33 +0100)]
Fix MSVC "result of 32-bit shift implicitly converted to 64 bits" warnings. NFCI.
Dmitri Gribenko [Tue, 20 Sep 2022 09:28:20 +0000 (11:28 +0200)]
Fix an unused warning in release build
David Spickett [Thu, 12 May 2022 15:38:43 +0000 (15:38 +0000)]
[lldb] Log when we cannot find an equivalent for a gdb register type
This happens if the type is described elsewhere in target xml as a
<flags> or <struct>.
Also hardcode the function names into the log messages because
if you use __FUNCTION__ in a lambda you just get "operator()".
Reviewed By: clayborg
Differential Revision: https://reviews.llvm.org/D134043
David Spickett [Mon, 1 Aug 2022 10:17:02 +0000 (10:17 +0000)]
[LLDB] Make instruction emulation context type private
This is the first step to being able to handle non
trivial types in the union.
info_type effects the lifetime of the objects in the union,
so making it private means we know you have to call one of the
Set<...> functions to change it.
Reviewed By: clayborg
Differential Revision: https://reviews.llvm.org/D134039
Dmitry Polukhin [Sat, 17 Sep 2022 15:30:30 +0000 (08:30 -0700)]
[clang] Update ReleaseNotes about a crash fix (Issue 53628)
Update ReleaseNotes about a crash fix (Issue 53628)
Test Plan: none
Differential Revision: https://reviews.llvm.org/D134112
David Spickett [Wed, 17 Aug 2022 13:58:28 +0000 (13:58 +0000)]
[LLDB] Format lldb-server's target XML
So that the XML isn't one giant line. Which wasn't
a problem for lldb but was for me trying to troubleshoot
it using the logs.
It now looks like:
```
<?xml version="1.0"?>
<target version="1.0">
<architecture>aarch64</architecture>
<feature>
<...>
<reg name="fpcr" .../>
</feature>
</target>
```
Reviewed By: labath
Differential Revision: https://reviews.llvm.org/D134035
Caroline Concatto [Wed, 31 Aug 2022 13:05:03 +0000 (14:05 +0100)]
[AArch64][NFC] Correctly rename mangling name for ldN.sret
Remove from the function name the predicate type and pointer type, because:
The predicate type in the name(nxvNi1) can be deduced from the overloaded
element count(nxvNEltTy).
The pointer type(p0EltTy) can be deduced from the overloaded element type.
Differential Revision: https://reviews.llvm.org/D133023
John McIver [Tue, 20 Sep 2022 08:39:40 +0000 (09:39 +0100)]
[Utils] Refactor update_cc_test_checks.py to use shutil
The package distutils is deprecated and removal is planned for Python 3.12.
All calls to distutils.spawn.find_executable are replaced with shutil.which.
Differential Revision: https://reviews.llvm.org/D134015
Jean Perier [Tue, 20 Sep 2022 08:39:39 +0000 (10:39 +0200)]
[flang] Deallocate WHERE masks after all assignments
Allocatable assignments were triggering lowering to clean-up
any WHERE mask temporaries, even if some assignments where left
in the WHERE construct.
This is because allocatable assignments lowering was being passed the
wrong statement context. Fix this by selecting the where/forall statement
context instead of a local one when there is one.
Differential Revision: https://reviews.llvm.org/D134197
eopXD [Tue, 20 Sep 2022 02:07:43 +0000 (19:07 -0700)]
[LSR] Fold terminating condition to other IV when possible
When the IV is only used by the terminating condition (say IV-A) and the loop
has a predictable back-edge count and we have another IV (say IV-B) that is an
affine add recursion, we will be able to calculate the terminating value of
IV-B in the loop pre-header. This patch adds attempts to replace IV-B as the
new terminating condition and remove IV-A. It is safe to do so since IV-A is
only used as the terminating condition.
This transformation is suitable to be appended after LSR as it may optimize the
loop into the situation mentioned above. The transformation can reduce number of
IV-s in the loop by one.
A cli option `lsr-term-fold` is added and default disabled.
Reviewed By: mcberg2021, craig.topper
Differential Revision: https://reviews.llvm.org/D132443
Jean Perier [Tue, 20 Sep 2022 08:15:15 +0000 (10:15 +0200)]
[flang] fix optional pointer TARGET argument lowering in ASSOCIATED
The TARGET argument of ASSOCIATED has a special lowering to deal with
POINTER and ALLOCATABLE optional actual arguments because they may be
dynamically absent. The previous code was doing a ternary
(mlir::SelectOp) to deal with this case, but generated invalid
code for the unused argument (loading a nullptr fir.ref<fir.box>). This
was not detected until D133779 was merged and modified how fir.load are
lowered to LLVM for fir.box types.
Replace the select by a proper if to prevent the fir.load from being
reachable in context where it should not.
Differential Revision: https://reviews.llvm.org/D134174
LLVM GN Syncbot [Tue, 20 Sep 2022 08:12:06 +0000 (08:12 +0000)]
[gn build] Port
55158efe1045
Nikolas Klauser [Tue, 30 Aug 2022 12:57:17 +0000 (14:57 +0200)]
[libc++] Remove MSVC code
It's been one and a half months now and nobody said anything, so I guess this code can be removed.
Reviewed By: ldionne, #libc
Spies: Mordante, libcxx-commits, mgorny, mstorsjo
Differential Revision: https://reviews.llvm.org/D132943
Nikolas Klauser [Mon, 15 Aug 2022 12:09:49 +0000 (14:09 +0200)]
[libc++][NFC] Refactor enable_ifs in vector
Using the `enable_if_t<..., int> = 0` style has the benefit that it works in all cases and makes function declarations easier to read because the function arguments and return type and SFINAE are separated. Unifying the style also makes it easier for people not super familiar with SFINAE to make sense of the code.
Reviewed By: Mordante, var-const, #libc, huixie90
Spies: huixie90, libcxx-commits
Differential Revision: https://reviews.llvm.org/D131868
Mateusz Mikuła [Tue, 20 Sep 2022 07:49:41 +0000 (10:49 +0300)]
[TargetLibraryInfo] Mark memrchr as unavailable on Windows
Otherwise LLVM will optimise strrchr into memrchr on Windows resulting in linker error:
```
$ cat memrchr_test.c
int main(int argc, char **argv) {
return (long)strrchr("KkMm", argv[argc-1][0]);
}
$ clang memrchr_test.c -O
memrchr_test.c:3:12: warning: cast to smaller integer type 'long' from 'char *' [-Wpointer-to-int-cast]
return (long)strrchr("KkMm", argv[argc-1][0]);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
ld.lld: error: undefined symbol: memrchr
>>> referenced by D:/msys64/tmp/memrchr_test-e7aabd.o:(main)
clang: error: linker command failed with exit code 1 (use -v to see invocation)
```
Example taken from MSYS2 Discord and tested with windows-gnu target.
Reviewed By: aeubanks
Differential Revision: https://reviews.llvm.org/D134134
Alvin Wong [Tue, 20 Sep 2022 07:49:01 +0000 (10:49 +0300)]
[LLD][COFF] Add load config checks to warn if incorrect for CFGuard
Control Flow Guard requires specific flags and VA's be included in the
load config directory to be functional. In case CFGuard is enabled via
linker flags, we can check to make sure this is the case and give the
user a warning if otherwise.
MSVC provides a proper `_load_config_used` by default, so this is more
relevant for the MinGW target in which current versions of mingw-w64
does not provide this symbol.
The checks (only if CFGuard is enabled) include:
- The `_load_config_used` struct shall exist.
- Alignment of the `_load_config_used` struct (shall be aligned to
pointer size.)
- The `_load_config_used` struct shall be large enough to contain the
required fields.
- The values of the following fields are checked against the expected
values:
- GuardCFFunctionTable
- GuardCFFunctionCount
- GuardFlags
- GuardAddressTakenIatEntryTable
- GuardAddressTakenIatEntryCount
- GuardLongJumpTargetTable
- GuardLongJumpTargetCount
- GuardEHContinuationTable
- GuardEHContinuationCount
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D133099
Alvin Wong [Tue, 20 Sep 2022 07:44:45 +0000 (10:44 +0300)]
[LLD][COFF] Improve symbol table info for import thunk
Import thunks themselves contain a jump or branch, which is code by
nature. Therefore the import thunk symbol should be marked as function
type in the symbol table to help with debugging.
The `__imp_` import symbol associated to the import thunk is also useful
for debugging. However, when the import symbol isn't directly referenced
outside of the import thunk, it doesn't normally get added to the symbol
table. This change teaches LLD to add the import symbol explicitly.
Reviewed By: mstorsjo
Differential Revision: https://reviews.llvm.org/D134169
Martin Storsjö [Sun, 18 Sep 2022 21:08:41 +0000 (00:08 +0300)]
[llvm-rc] [test] Fix the windres-preproc test with other mingw triple forms. NFC.
When invoking the preprocessor, llvm-windres produces a mingw
triple; if the llvm default target triple is a mingw target, it
is used as is - if not, a mingw style triple is generated.
Relax the testcase, to not strictly require "w64" as vendor (allow
the vendor field to be missing entirely), and allow both
"mingw32" and "windows-gnu" as OS/environment (where both forms are
common, while the latter is the normalized form).
Differential Revision: https://reviews.llvm.org/D134148
Michał Górny [Mon, 19 Sep 2022 18:23:44 +0000 (20:23 +0200)]
[clang] [Driver] Do not transform explicit --config filename
Disable transformations (e.g. attempting to replace target architecture)
in the config filename that is passed explicitly via `--config`. This
behavior is surprising and confusing -- if user passes an explicit
config filename, Clang should use it as is. The transformations are
still applied when the name is deduced from filename.
Update the tests accordingly. This primarily ensures that full filename
with .cfg suffix is passed to --config (appending `.cfg` implicitly is
not documented, and would collide with use of filenames with other
suffixes). The config-file2.c suite is removed entirely as it tested
the transformations on the argument to --config. However, the aspects
of that that were not tested as part of config-file3.c are now added
there (based on config filename deduced from executable).
This change streamlines the code in Driver::loadConfigFile(), opening
the possibility of further changes, including support for handling
multiple --config options and refactoring of filename deduction.
Differential Revision: https://reviews.llvm.org/D134208
wlei [Tue, 20 Sep 2022 05:30:24 +0000 (22:30 -0700)]
remove the internal signatures from perf binaries
Jonas Devlieghere [Mon, 19 Sep 2022 23:20:50 +0000 (16:20 -0700)]
[lldb] Actually support more than 32 logging categories
In January, Greg put up a patch (D117382) to support, among other
things, more than 32 log categories. That led to a bunch of nice
cleanups, but categories remained constrained because different parts of
the code were still using uint32_t. This patch fixes the remaining
issues and makes it possible to add a 32nd log category.
Differential revision: https://reviews.llvm.org/D134245
Serge Pavlov [Mon, 19 Sep 2022 07:14:00 +0000 (14:14 +0700)]
[X86][GlobalISel] Add support for sret demotion
The change add support for the cases when return value is passed in
memory rathen than in registers.
Differential Revision: https://reviews.llvm.org/D134181
Jonathon Penix [Fri, 2 Sep 2022 03:04:18 +0000 (20:04 -0700)]
[flang] Give explicit convert= specifiers precedence over FORT_CONVERT
Currently, the FORT_CONVERT environment variable has the highest priority when
setting the endianness conversion for unformatted files. In discussing the
appropriate priority for the fconvert option, convert specifiers were decided
to take highest priority.
This patch also initializes the open statement convert state to unknown
to disambiguate cases where the convert specifier was not provided from
cases where convert=native was set. This makes it possible to defer to the
environment setting where appropriate.
Differential Revision: https://reviews.llvm.org/D133237
luxufan [Mon, 19 Sep 2022 02:02:49 +0000 (02:02 +0000)]
[MemorySSA] Add test that all incoming values of phi node could be translated
Differential Revision: https://reviews.llvm.org/D134160
luxufan [Mon, 19 Sep 2022 02:24:10 +0000 (02:24 +0000)]
[MemorySSA][NFC] Use const whenever possible
Differential Revision: https://reviews.llvm.org/D134162
eopXD [Tue, 23 Aug 2022 08:44:15 +0000 (01:44 -0700)]
[LSR] Precommit test for D132443
Pre-commit test for D132443
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D132452
Ron Lieberman [Tue, 20 Sep 2022 01:30:34 +0000 (01:30 +0000)]
revert
684f76643 [Clang][OpenMP] Codegen generation for has_device_addr claues.
breaks amdgpu buildbot
Phoebe Wang [Tue, 20 Sep 2022 00:44:11 +0000 (08:44 +0800)]
[X86][fastcall][vectorcall] Move capability check before free register update
When passing arguments with `__fastcall` or `__vectorcall` in 32-bit MSVC, the following arguments have chance to be passed by register if the current one failed. `__regcall` from ICC is on the contrary: https://godbolt.org/z/4MPbzhaMG
All the three calling conversions are not supported in GCC.
Fixes: #57737
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D133920
Vitaly Buka [Mon, 12 Sep 2022 05:00:37 +0000 (22:00 -0700)]
[msan] Handle shadow of masked instruction
Origin handling is not implemented yet.
Reviewed By: kda
Differential Revision: https://reviews.llvm.org/D133682
Jonas Devlieghere [Tue, 20 Sep 2022 00:36:19 +0000 (17:36 -0700)]
[lldb] Appease the MSCV compiler
Fix error C2027: use of undefined type 'llvm::MemoryBuffer'.
Jennifer Yu [Tue, 20 Sep 2022 00:04:09 +0000 (17:04 -0700)]
Fix test case which is not working for AMDGPU.
This is for the change of
Differential Revision: https://reviews.llvm.org/D134186
Matt Arsenault [Mon, 19 Sep 2022 23:08:10 +0000 (19:08 -0400)]
AMDGPU: Add some tests for atomics with excess alignment
Matt Arsenault [Mon, 19 Sep 2022 22:09:39 +0000 (18:09 -0400)]
VectorCombine: Pass through AssumptionCache