Luke Lau [Thu, 9 Mar 2023 14:35:57 +0000 (14:35 +0000)]
[RISCV] Model interleave and deinterleave shuffles in cost model
Interleave and deinterleave shuffles are lowered by a more efficient
sequence if the element size is smaller than ELEN.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D145678
Kirill Stoimenov [Fri, 10 Mar 2023 01:07:43 +0000 (01:07 +0000)]
[HWASAN][LSAN] Fix buildbot failure.
Paul Walker [Fri, 10 Mar 2023 00:39:39 +0000 (00:39 +0000)]
[SVE] Precommit multiuse predicated compare tests.
Kirill Stoimenov [Mon, 6 Mar 2023 23:55:32 +0000 (23:55 +0000)]
[HWASAN][LSAN] Disable tests which don't pass in HWASAN+LSAN mode
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D145727
Michael Buch [Wed, 8 Mar 2023 11:32:06 +0000 (11:32 +0000)]
[lldb][InstrumentationRuntime] Make 'data' struct anonymous in order to avoid collisions with types in the debuggee
The `UBSAN`/`ASAN` plugins would previously call the internal helper
structure injected into the source expression `struct data { ... };`
This occasionally collided with user-defined types of the same (quite
common) name. In the presence of varibale with the same name LLDB would
simply fail to run the injected `InstrumentationRuntime` expressions
with:
```
warning: cannot evaluate AddressSanitizer expression:
expression failed to parse:
error: <user expression 2>:2:5: must use 'struct' tag to refer to type 'data' in this scope
data t;
```
In the presence of another type called 'data', LLDB would most likely
crash with something like:
```
0x00000001198de614 clang::ASTNodeImporter::ImportDeclContext(clang::DeclContext*, bool) + 1088
0x00000001198de614 clang::ASTNodeImporter::ImportDeclContext(clang::DeclContext*, bool) + 1088
0x0000000119907930 clang::ASTImporter::ImportDefinition(clang::Decl*) + 596
0x00000001163db928 lldb_private::ClangASTImporter::ASTImporterDelegate::ImportDefinitionTo(clang::Decl*, clang::Decl*) + 100
0x00000001163da070 (anonymous namespace)::CompleteTagDeclsScope::~CompleteTagDeclsScope() + 572
...
```
...because it got the types confused.
This patch makes these structures anonymous so there's no
chance of clashing with other types in the program. This is
already the approach taken in `UBSan/InstrumentationRuntimeABSan.cpp`.
**Testing**
- API tests still pass
- Tested manually that the `memory history` command now works in the presence of a local called `data`
Differential Revision: https://reviews.llvm.org/D145569
Jun Zhang [Fri, 10 Mar 2023 00:00:41 +0000 (08:00 +0800)]
[InstCombine] Reduce absolute diff from min+max+sub
This patch implements fold: max(a,b) nsw/nuw - min(a,b) --> abs(a nsw - b)
Alive2: https://alive2.llvm.org/ce/z/4yLp7D
Fixes: https://github.com/llvm/llvm-project/issues/61228
Signed-off-by: Jun Zhang <jun@junz.org>
Differential Revision: https://reviews.llvm.org/D145540
Arthur Eubanks [Thu, 9 Mar 2023 23:42:27 +0000 (15:42 -0800)]
[llvm-c] Remove pointee support from LLVMGetElementType
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D145717
Ben Shi [Wed, 8 Mar 2023 10:03:49 +0000 (18:03 +0800)]
[RISCV][NFC] Add tests for SLP vectorization of math functions
RISCV has "vfabs.v" and "vfsqrt.v" so math functions abs and sqrt
can be SLP vectorized. But others exp/log/sin/asin/sinh/asinh/...
can not.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D145562
Renaud-K [Thu, 9 Mar 2023 02:39:40 +0000 (18:39 -0800)]
Break circular dependency between FIR dialect and utilities
Emilio Cota [Thu, 9 Mar 2023 22:08:00 +0000 (17:08 -0500)]
[mlir][vector] remove unnecessary VectorTransformOps include
While at it, add a dep that we missed in https://reviews.llvm.org/D145638.
Reviewed By: kuhar, dcaballe
Differential Revision: https://reviews.llvm.org/D145731
Jason Molenda [Thu, 9 Mar 2023 22:53:14 +0000 (14:53 -0800)]
Slight refinement to a change yesterday in metadata-added binaries
When ObjectFileMachO::LoadCoreFileImages load a binary into the
target with a valid load address, we don't need to re-load its
segments into the Target's SectionLoadList again. But we should
still call ModulesDidLoad on these modules so breakpoints can be
inserted etc.
Bruno Cardoso Lopes [Tue, 7 Mar 2023 20:51:34 +0000 (12:51 -0800)]
[Coroutines] Fix premature conversion of return object
Fix https://github.com/llvm/llvm-project/issues/56532
Effectively, this reverts behavior introduced in https://reviews.llvm.org/D117087,
which did two things:
1. Change delayed to early conversion of return object.
2. Introduced RVO possibilities because of early conversion.
This patches fixes (1) and removes (2). I already worked on a follow up for (2)
in a separated patch. I believe it's important to split these two because if the RVO
causes any problems we can explore reverting (2) while maintaining (1).
Notes on some testcase changes:
- `pr59221.cpp` changed to `-O1` so we can check that the front-end honors
the value checked for. Sounds like `-O3` without RVO is more likely
to work with LLVM optimizations...
- Comment out delete members `coroutine-no-move-ctor.cpp` since behavior
now requires copies again.
Differential Revision: https://reviews.llvm.org/D145639
Christopher Ferris [Thu, 9 Mar 2023 00:56:24 +0000 (16:56 -0800)]
[scudo] Add a fast get time version.
On Android, the _COARSE version of clock_gettime is about twice as fast.
Therefore, add a getMonotonicTimeFast function that is used in the
releaseToOSMaybe functions.
Reviewed By: Chia-hungDuan
Differential Revision: https://reviews.llvm.org/D145636
bzcheeseman [Thu, 9 Mar 2023 19:02:21 +0000 (11:02 -0800)]
[ORC] Fix name shadowing issue, NFC.
Testing the COFFPlatform on MSVC, a name shadowing issue surfaced where `LoadDynLibrary` inside the constructor was actually using the moved-from function argument. This patch simply renames the argument to avoid that shadowing.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D145710
Aart Bik [Thu, 9 Mar 2023 18:44:46 +0000 (10:44 -0800)]
[sparse][mlir] test transposition on sorted COO
DO NOT SUBMIT YET, test exposes bug
Reviewed By: Peiming
Differential Revision: https://reviews.llvm.org/D145708
Peiming Liu [Wed, 8 Mar 2023 23:24:16 +0000 (23:24 +0000)]
[mlir][sparse] deduplicate non-unique coordinates unconditionally
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D145621
Leonard Chan [Thu, 9 Mar 2023 21:42:04 +0000 (21:42 +0000)]
[hwasan] Suppress leaks from dlsym with hwasan+lsan just like with
asan+lsan
We should follow suite with how asan handles this now that lsan also
works with hwasan.
Differential Revision: https://reviews.llvm.org/D145613
Nikolas Klauser [Thu, 9 Mar 2023 21:35:40 +0000 (22:35 +0100)]
[libc++] Fix include order in <optional>
Nadeem, Usman [Wed, 1 Feb 2023 23:09:23 +0000 (15:09 -0800)]
[Flang] Add support to use LTO specific pipelines
Thin and full LTO modes use different pre-link pipelines compared to
regular compilation. This patch adds support for calling those pipelines.
This patch closely mimics Clang's implementation with the exception that I
changed the codegen option name from `PrepareForLTO` to `PrepareForFullLTO`
to be more precise.
With this patch:
- Compilation for full LTO should be as we expect (except possibly
missing optimizations enabled by module summaries which we do not
produce yet).
- thinLTO uses the correct prelink pipeline but will use the postlink
backend for fullLTO due to missing metadata and summary in the llvm
module. I have added a warning regarding this: `flang-new: warning: the
option '-flto=thin' is a work in progress`.
Differential Revision: https://reviews.llvm.org/D142420
Change-Id: I6b94b775b5b8e93340e520c5cd4bf60834b2e209
Jez Ng [Wed, 8 Mar 2023 04:48:25 +0000 (20:48 -0800)]
[LTO] Add debug logging for module ID <-> path mapping
It's helpful
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D145541
Peiming Liu [Thu, 9 Mar 2023 21:22:59 +0000 (21:22 +0000)]
[mlir][sparse] add a check test for foreach operation on constant sparse tensor
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D145728
Arthur Eubanks [Thu, 9 Mar 2023 19:27:52 +0000 (11:27 -0800)]
[docs] Mention LLVM_USE_LINKER/LLVM_PARALLEL_{COMPILE,LINK}_JOBS in GettingStarted
As requested in D145413.
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D145714
Peiming Liu [Thu, 9 Mar 2023 20:46:04 +0000 (20:46 +0000)]
[mlir][sparse] fix bugs when convert coo to coo but with different dim ordering
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D145723
Joshua Batista [Thu, 9 Mar 2023 19:31:31 +0000 (11:31 -0800)]
Add codegen for llvm exp/exp2 elementwise builtins
Add codegen for llvm exp/exp2 elementwise builtin
The exp/exp2 elementwise builtins are necessary for HLSL codegen.
Tests were added to make sure that the expected errors are encountered when these functions are given inputs of incompatible types.
The new builtins are restricted to floating point types only.
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D145270
Jakub Kuderski [Thu, 9 Mar 2023 20:13:28 +0000 (15:13 -0500)]
[mlir][vector] Add missing build dependencies for VectorToGPU
Post-commit fix for https://reviews.llvm.org/D145638.
Jakub Kuderski [Thu, 9 Mar 2023 19:56:20 +0000 (14:56 -0500)]
[mlir][vector][nvgpu] Move MMA contraction preparation to VectorUtils
This pattern is not specific to nvgpu; I intend to use in SPIR-V codegen. `VectorTransforms` seems like a more generally useful place.
In addition:
- Fix a bug in the second condition (the dimensions were swapped for RHS).
- Add tests.
- Add support for externally provided filter functions, similar to other vector transforms.
- Prefer to transpose before zero/sign-extending inputs.
Reviewed By: ThomasRaoux
Differential Revision: https://reviews.llvm.org/D145638
Snehasish Kumar [Thu, 9 Mar 2023 03:55:39 +0000 (03:55 +0000)]
[memprof] Add scripts to automate testdata regeneration.
The memprof profiles and binaries need to be updated in case of version
updates. This change adds three scripts for llvm-profdata, clang and
llvm tests where memprof profiles are used as inputs. Also update the
tests, profiles and binaries in this change. Change based on the review
suggestions in D145023.
Reviewed By: tejohnson
Differential Revision: https://reviews.llvm.org/D145644
Jonas Devlieghere [Thu, 9 Mar 2023 19:32:17 +0000 (11:32 -0800)]
[lldb] Skip TestSymbolFileJSON on Windows
Disable the test as Windows (or at least the bot) doesn't have 'strip':
'strip' is not recognized as an internal or external command
Jennifer Yu [Thu, 9 Mar 2023 16:14:56 +0000 (08:14 -0800)]
Revert "Revert "Add map info for dereference pointer.""
This reverts commit
8cf85a0cadb033fed3d96aa5283deb4bfbbaf2c8.
This is add back change of "Add map info for dereference pointer."
In addition turn off test run on amdgpu, since I don't know the way to
reprodue the problem.
Fehr Mathieu [Thu, 9 Mar 2023 18:27:50 +0000 (18:27 +0000)]
[mlir] Add setters for Dynamic[Type|Attr]Definition
Add functions to set the verifier, printer, and parser of
dynamic attributes definitions, and dynamic type definitions.
This feature was already implemented for dynamic operations, but
is missing for attributes and types. This is necessary to define
attributes and types verifiers that refer to each others in a cyclic way.
Differential Revision: https://reviews.llvm.org/D144690
Michael Jones [Thu, 9 Mar 2023 00:39:45 +0000 (16:39 -0800)]
[libc][docs] update list of completed stdio funcs
Reviewed By: lntue
Differential Revision: https://reviews.llvm.org/D145633
Jonas Devlieghere [Thu, 9 Mar 2023 18:26:21 +0000 (10:26 -0800)]
[lldb] Add nullptr check to SymbolVendorWasm
Add the same nullptr check to SymbolVendorWasm that was added to
SymbolVendorELF.
Jonas Devlieghere [Thu, 9 Mar 2023 18:20:41 +0000 (10:20 -0800)]
[dsymutil] Fix offset calculation for universal binaries
The Mach-O file format uses 32-bit values to encodes offsets which they
cannot exceed UIN32_MAX (4GB). The Mach-O file itself can be larger than
4GB as long as none of the offsets fall within this limit.
For universal binaries, dsymutil determines if the offset is going to
exceed the 4GB limit by computing the size of the header and adding it
to the size of all the slices. This is incorrect because it computes the
end offset of the final slice. For the purpose of the 4GB limit, only
the starting offset matters. The size of the last slice is irrelevant as
long as it itself is a valid Mach-O.
rdar://
104435018
Differential revision: https://reviews.llvm.org/D145637
Michal Paszkowski [Thu, 9 Mar 2023 16:22:33 +0000 (17:22 +0100)]
[SPIR-V] Report modifying IR in SPIRVPrepareFunctions
This change fixes "Pass modifies its input and doesn't
report it" error when running SPIRVPrepareFunctions
pass with LLVM_ENABLE_EXPENSIVE_CHECKS enabled.
Differential Revision: https://reviews.llvm.org/D145121
Zain Jaffal [Thu, 9 Mar 2023 18:16:19 +0000 (18:16 +0000)]
[ConstraintElimination] Fix undefined behaviour in shl decomposition
Add checks to prevent decomposing constants bigger than 64.
relates to https://github.com/llvm/llvm-project/issues/61127
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D145677
Peiming Liu [Tue, 7 Mar 2023 21:52:03 +0000 (21:52 +0000)]
[mlir][sparse] deduplicate non-unique coordinates when coiterating collapsed COO tensors.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D145532
Hanhan Wang [Thu, 9 Mar 2023 18:03:31 +0000 (10:03 -0800)]
[mlir][tensor] Expose padding requirement of pack ops to a static method
It also simplifies the implementation of the method. The map is not needed in the check.
Reviewed By: chelini
Differential Revision: https://reviews.llvm.org/D145522
Ed Maste [Wed, 14 Dec 2022 21:03:56 +0000 (16:03 -0500)]
[libc++] Add CI job testing on FreeBSD
Reviewed by: ldionne, Mordante
Differential Revision: https://reviews.llvm.org/D128084
Peter Klausler [Fri, 3 Mar 2023 17:52:41 +0000 (09:52 -0800)]
[flang] Ensure "dtv" args passed by descriptor when polymorphic
The predicate used to determine whether the initial "dtv" argument
to a user-defined derived type I/O procedure specified by a generic
interface that is not in the derived type definition is wrong.
The "dtv" argument must be passed via a descriptor whenever it is
polymorphic.
Differential Revision: https://reviews.llvm.org/D145255
Peter Klausler [Fri, 3 Mar 2023 16:17:31 +0000 (08:17 -0800)]
[flang] Ensure that CLASS(*) component descriptors have addenda
In the calculation of derived type component byte sizes, ensure
that CLASS(*) unlimited polymorphic components have space allocated
for their addenda.
Differential Revision: https://reviews.llvm.org/D145248
Luke Lau [Thu, 9 Mar 2023 16:14:25 +0000 (16:14 +0000)]
[RISCV][NFC] Add tests for interleaved accesses in loop vectorizer
Precommit test for D145155
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D145697
Paul Robinson [Thu, 9 Mar 2023 17:26:17 +0000 (09:26 -0800)]
[PS4/PS5] Remove C_INCLUDE_DIRS handling
We don't configure this way and don't intend to.
Alexey Lapshin [Tue, 7 Mar 2023 12:46:21 +0000 (13:46 +0100)]
[DebugInfo][DWARF] Add doesFormBelongToClass function.
The result of DWARFFormValue::isFormClass depends on DWARF version in some cases.
The current implementation takes DWARF version from the stored DWARFUnit.
If there is no stored DWARFUnit then the current behavior is to assume
DwarfVersion <= 3. This patch adds new function which has a DWARF version as a
parameter so it is possible to check form class for various DWARF versions.
Differential Revision: https://reviews.llvm.org/D145499
Aiden Grossman [Thu, 9 Mar 2023 17:20:18 +0000 (17:20 +0000)]
[ELF] Fix build after getSectionAndRelocations Refactoring
Didn't fail locally for some reason with my gcc toolchain.
Erich Keane [Wed, 8 Mar 2023 20:12:33 +0000 (12:12 -0800)]
Revert two patches to fix GH58452 regression
GH58452 is a regression in the 16.0 release branch caused by both:
b8a1b698afb2fc84819c7596090aabf4d826b436 and
3a0309c53674be56b5cfce038d78a0c2c6e2a98c
This patch reverts both of those to make the 'valid' code stop
diagnosing
at the expense of crashes on invalid + unclear diagnostics.
This patch also adds the tests from GH58452 to prevent any
re-application from breaking this again.
Revert "[clang] Improve diagnostics for expansion length mismatch"
This reverts commit
3a0309c53674be56b5cfce038d78a0c2c6e2a98c.
Revert "[clang] fix missing initialization of original number of expansions"
This reverts commit
b8a1b698afb2fc84819c7596090aabf4d826b436.
Differential Revision: https://reviews.llvm.org/D145605
Aiden Grossman [Sat, 25 Feb 2023 10:18:34 +0000 (10:18 +0000)]
[ELF] Move getSectionAndRelocations to ELF.cpp from ELFDumper.cpp
This refactoring will allow for this utility function to be used in
other places in the codebase outside of the llvm-readobj tool.
Reviewed By: jhenderson, rahmanl
Differential Revision: https://reviews.llvm.org/D144783
Siva Chandra [Thu, 9 Mar 2023 16:53:31 +0000 (16:53 +0000)]
[libc] Extend the linux thread implementation for riscv64.
Also, added riscv64 startup code for static linking which is used
by the integration tests. Functions from the C standard threads
library have been enabled.
Reviewed By: mikhail.ramalho
Differential Revision: https://reviews.llvm.org/D145670
Michael Kruse [Thu, 9 Mar 2023 16:48:02 +0000 (10:48 -0600)]
[Polly] Fix plugin build.
The target_link_libraries call was accidentially removed in
D142580 (Remove Polly-ACC).
Florian Hahn [Thu, 9 Mar 2023 16:52:22 +0000 (17:52 +0100)]
[VPlan] Replace invariance fields from VPWidenGEPRecipe.
There is no need to store information about invariance in the recipe.
Replace the fields with checks of the operands using
isDefinedOutsideVectorRegions.
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D144487
Philip Reames [Thu, 9 Mar 2023 16:36:45 +0000 (08:36 -0800)]
[ASAN] Address a style issue noticed during review of D145175 [nfc]
Valentin Clement [Thu, 9 Mar 2023 16:48:11 +0000 (17:48 +0100)]
[flang] Check if dummy is allocated before deallocation for optional intent(out)
Similary to non-optional argument, check if the argument is allocated before
doing the deallocation for intent(out) optional.
Depends on D145679
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D145682
Luke Lau [Thu, 9 Mar 2023 14:33:35 +0000 (14:33 +0000)]
[RISCV][NFC] Add tests for interleave shuffle costs
Pre-commit test for D145678
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D145689
Chia-hung Duan [Wed, 8 Mar 2023 23:08:50 +0000 (23:08 +0000)]
[scudo] Slightly improve the handling of last block in a region
Instead of going through all those trailing blocks, just count the
number and increase the counter at once.
Reviewed By: cferris
Differential Revision: https://reviews.llvm.org/D145419
Philip Reames [Thu, 9 Mar 2023 15:55:07 +0000 (07:55 -0800)]
[ASAN] Support memory checks on scalable vector typed loads and stores
This only covers the common load/store case. There will be further patches required for masked load/store and some of the fast-path optimization cases.
Differential Revision: https://reviews.llvm.org/D145175
zhongyunde [Thu, 9 Mar 2023 15:48:51 +0000 (23:48 +0800)]
[MergeICmps] Adapt to non-eq comparisons, fix bug for cases need be spilted
Fix the last runtime issue as some sequent comparisons need be spilted.
For the origin equal comparisons chain, the new spilted Icmp chain will
still be end with equal, while for the new not-equal comparisons chain,
the new spilted Icmp chain will still be end with equal, so should address
this carefully, see detail wih case partial_sequent_ne.
Thanks for @glandium and @ayzhao report the runtime issue and carefully
examine.
Fix https://github.com/llvm/llvm-project/issues/59740.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D141188
Timm Bäder [Mon, 6 Mar 2023 13:57:19 +0000 (14:57 +0100)]
[clang][Interp][NFC] Take a const Descriptor* in dtor,move,ctorFns
We are not mutating the descriptors here.
Valentin Clement [Thu, 9 Mar 2023 15:05:57 +0000 (16:05 +0100)]
[flang] Only apply PointerIsAssociatedWith fix for derived-type
Only apply the change made in D145604 for derived-type.
zero-sized character for example are supposed to give a false result
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D145675
Valentin Clement [Thu, 9 Mar 2023 14:47:59 +0000 (15:47 +0100)]
[flang] Reset dynamic type for optional intent(out) allocatable polymorphic dummy
Allocatable intent(out) are deallocated at the beginning of a function/subroutine.
For polyrmophic entities, the dynamic type need to be reseted to the declared
type. This patch makes sure this is done when the dummy argument is optional and
present.
Depends on D145674
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D145679
Valentin Clement [Thu, 9 Mar 2023 14:46:59 +0000 (15:46 +0100)]
[flang] Simplify same_type_as condition
Restore the behavior changed in D145384 and add proper
unit tests.
Unallocated unlimited poymorphic allocatable and disassociated
unlimited polymorphic pointer should return false.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D145674
ManuelJBrito [Thu, 9 Mar 2023 14:31:33 +0000 (14:31 +0000)]
[X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS
Ignoring freeze(undef) if it has multiple uses in LowerAVXCONCAT_VECTORS
causes the custom INSERT_SUBVECTOR for vector widening to be ignored.
Differential Revision: https://reviews.llvm.org/D144903
Aaron Ballman [Thu, 9 Mar 2023 14:26:13 +0000 (09:26 -0500)]
Switch from int to unsigned int; NFC
This silences a GCC conversion diagnostic about assigning `1` to a
1-bit signed bit-field changing the value from `1` to `-1`.
Co-authored-by: Igor Kushnir <igorkuo@gmail.com>
Valery Pykhtin [Tue, 28 Feb 2023 18:00:48 +0000 (19:00 +0100)]
[AMDGPU] Speedup GCNDownwardRPTracker::advanceBeforeNext
The function makes liveness tests for the entire live register set for every instruction it passes by.
This becomes very slow on high RP regions such as ASAN enabled code.
Instead only uses of last tracked instruction should be tested and this greatly improves compilation time.
This patch revealed few bugs in SIFormMemoryClauses and PreRARematStage::sinkTriviallyRematInsts which should
be fixed first.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D136267
Kiran Chandramohan [Thu, 9 Mar 2023 14:05:08 +0000 (14:05 +0000)]
[Flang][OpenMP] Restrict check to worksharing construct reductions
The outer context private check for reduction variables was firing
for all constructs. This check is not applicable to non-worksharing
constructs.
OpenMP 5.2: Section 5.5.8
A list item that appears in a reduction clause on a worksharing construct
must be shared in the parallel region to which a correspodning worksharing
region binds.
Reviewed By: peixin
Differential Revision: https://reviews.llvm.org/D144824
Diana Picus [Thu, 9 Mar 2023 13:50:14 +0000 (14:50 +0100)]
AMDGPU: Update checks for a couple of tests. NFC
Peter Steinfeld [Mon, 6 Mar 2023 21:00:10 +0000 (13:00 -0800)]
[Flang] Fix ALLOCATE with MOLD where MOLD is a scalar
We were failing tests where an ALLOCATE statement that allocated an
array had a non-character scalar MOLD argument.
I fixed this by merging the code for ALLOCATE statements with MOLD and
SOURCE arguments.
Differential Revision: https://reviews.llvm.org/D145418
Dylan Fleming [Mon, 6 Mar 2023 15:47:16 +0000 (15:47 +0000)]
[Flang][MLIR][OpenMP] Add support for logical eqv in worksharing-loop
The patch adds the lowering from Flang parse-tree to FIR+OpenMP. The
conversion code is also added in MLIR.
Reviewed By: kiranchandramohan
Differential Revision: https://reviews.llvm.org/D133442
Co-authored-by: Kiran Chandramohan <kiran.chandramohan@arm.com>
Joseph Huber [Thu, 9 Mar 2023 13:53:47 +0000 (07:53 -0600)]
[Libomptarget] Remove unused arguments from bitcode compilation
Summary:
We passed `-fopenmp-target=` when we compiled the bitcode, which isn't
necessary since the 15 release. Also adjust an error message.
Krasimir Georgiev [Thu, 9 Mar 2023 13:02:45 +0000 (14:02 +0100)]
[libc++] Add missing includes to optional
After commit
0a4aa8a122aa097499c498b639a75b5e9a73e9f0 we see failures
like:
optional:722:7: error: missing '#include <__type_traits/conjunction.h>';
'_And' must be declared before it is used
and:
optional:683:46: error: missing '#include
<__type_traits/disjunction.h>'; '_Or' must be declared before it is used
using __check_constructible_from_opt = _Or<
Adding these here fixes that.
I'm not familiar with the libcxx codebase, will ask the author to take a
look too.
Ed Maste [Wed, 1 Mar 2023 22:13:09 +0000 (17:13 -0500)]
[libc++] Remove XFAIL for a test that now passes on FreeBSD
Aaron Ballman [Thu, 9 Mar 2023 13:02:31 +0000 (08:02 -0500)]
Revert "Add __builtin_set_flt_rounds"
This reverts commit
24b823554acd25009731b2519880aa18c7263550.
These changes broke the PPC build bot.
Aaron Ballman [Thu, 9 Mar 2023 13:01:55 +0000 (08:01 -0500)]
Revert "Disable test for __builtin_set_flt_rounds to avoid breaking PPC buildbot"
This reverts commit
d4fcc692ee15b2c6d249daabe31208d6a5afa025.
(This change removed all test coverage from a previous change, also to be reverted.)
Joseph Huber [Thu, 9 Mar 2023 12:45:50 +0000 (06:45 -0600)]
[libc] Add missing header for LIBC_INLINE macro
Summray:
Forgot to do this in the last patch
Joseph Huber [Wed, 8 Mar 2023 21:17:14 +0000 (15:17 -0600)]
[libc] Add intrinsic for thread fence to the atomic support
This function mimics the std::atomic_thread_fence function from
<atomic>. This has no uses in source currently, but this will be used by
the proposed RPC client for the GPU mode support. There is varying
support for direct memory ordering for the GPU atomics on shared memory
resources. So the implementation will use relaxed atomics and explicit
memory fences.
Some additional work may need to be done to map this to `NVPTX` system
level fences.
Reviewed By: sivachandra
Differential Revision: https://reviews.llvm.org/D145608
Simon Pilgrim [Thu, 9 Mar 2023 12:32:11 +0000 (12:32 +0000)]
[X86] combineConcatVectorOps - add basic TRUNCATE handling on AVX512 targets
This really should be in shuffle combining, but we're still struggling to handle mismatched vector sizes
Advenam Tacet [Thu, 9 Mar 2023 12:20:28 +0000 (13:20 +0100)]
[libc++] Safe allocator tests
This revision adds:
- New test allocator, which cleans memory during allocation and deallocation,
- tests using that allocator to vector.
This patch is part of our efforts to add support for ASan annotations with every
allocator.
This commit adds a new allocator for testing purposes only. The safe allocator
ensures that memory is cleand (zeroed) during allocation and deallocation, and
is intendted to test ASan annotations for every allocator in std::vector.
Check: D136765
Those tests should work correctly, even if support for every allocator in std::vector
is not yet available.
Support in ASan API was added here: rGdd1b7b797a116eed588fd752fbe61d34deeb24e4
Reviewed By: philnik, #libc
Spies: libcxx-commits
Differential Revision: https://reviews.llvm.org/D145597
Sacha Ballantyne [Thu, 9 Mar 2023 10:22:05 +0000 (10:22 +0000)]
[Flang] Give fir.if RegionBranchOpInterface
fir.if currently isn't treated as a 'proper' conditional, so passes are unable to determine which regions are executed at times.
This patch gives fir.if this interface, which shouldn't do too much on its own but should allow future changes to take advantage
for various purposes
Reviewed By: vzakhari
Differential Revision: https://reviews.llvm.org/D145165
Simon Pilgrim [Thu, 9 Mar 2023 12:07:40 +0000 (12:07 +0000)]
[X86] Add test case showing failure to fold concat_v32i8(trunc_v16i16(x),trunc_v16i16(y)) -> trunc_v32i16(concat_v16i16(x,y))
We have similar tests but nothing that should only fold on BWI targets
Nikita Popov [Thu, 9 Mar 2023 11:41:40 +0000 (12:41 +0100)]
[ConstantFolding] Preserve inbounds when casting GEP indices
This canonicalization just makes the implicit sext/trunc explicit,
and does not affect the inbounds-ness of the GEP.
Nikita Popov [Thu, 9 Mar 2023 12:01:50 +0000 (13:01 +0100)]
[InstSimplify] Fix newlines in test (NFC)
This file used CRLF newlines.
ManuelJBrito [Thu, 9 Mar 2023 11:55:09 +0000 (11:55 +0000)]
Revert "[X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS"
This reverts commit
e2817933fdd4417673eb7f96a2842ae2dca8fc59.
Nikita Popov [Thu, 9 Mar 2023 11:30:48 +0000 (12:30 +0100)]
Revert "[clang-format] Add -j to clang-format-diff to speed up formatting"
This reverts commit
d14e7ee3d17cfa60d44256d742c10e9949a6048f.
Makes clang_format_diff.py always fail with:
Traceback (most recent call last):
File "/home/npopov/repos/llvm-project/clang/tools/clang-format/clang-format-diff.py", line 177, in <module>
main()
File "/home/npopov/repos/llvm-project/clang/tools/clang-format/clang-format-diff.py", line 174, in main
process_subprocess_result(proc, args)
File "/home/npopov/repos/llvm-project/clang/tools/clang-format/clang-format-diff.py", line 42, in process_subprocess_result
with open(filename) as f:
^^^^^^^^
NameError: name 'filename' is not defined
Markus Mützel [Thu, 9 Mar 2023 10:55:53 +0000 (12:55 +0200)]
Reapply [lld][flang] Add exceptions for Flang runtime libraries on MinGW.
When linking a shared library with Flang on MinGW, the functions from the
Flang runtime are exported from the shared library. When trying to link an
executable to that library using Flang, the linker errors out because the
functions from the runtime conflict with the functions exported from the
shared library.
Add the Flang runtime libraries to the list of libraries for which no
symbols are exported.
Reapplying the patch with the git author name corrected.
Reviewed By: mstorsjo
Differential Revision: https://reviews.llvm.org/D145389
Martin Storsjö [Thu, 9 Mar 2023 11:18:04 +0000 (13:18 +0200)]
Revert "[lld][flang] Add exceptions for Flang runtime libraries on MinGW."
This reverts commit
04fc67bf08daef5031f78c4a52a069ff29cff099.
The patch was applied with the wrong git author name (forgot to
amend in the right user name after applying the patch); I'll
reapply it with the right author name.
Simon Pilgrim [Thu, 9 Mar 2023 11:16:44 +0000 (11:16 +0000)]
[X86] combineConcatVectorOps - add ADD/SUB/MUL handling
Martin Storsjö [Thu, 9 Mar 2023 10:55:53 +0000 (12:55 +0200)]
[lld][flang] Add exceptions for Flang runtime libraries on MinGW.
When linking a shared library with Flang on MinGW, the functions from the
Flang runtime are exported from the shared library. When trying to link an
executable to that library using Flang, the linker errors out because the
functions from the runtime conflict with the functions exported from the
shared library.
Add the Flang runtime libraries to the list of libraries for which no
symbols are exported.
Reviewed By: mstorsjo
Differential Revision: https://reviews.llvm.org/D145389
ManuelJBrito [Thu, 9 Mar 2023 10:57:05 +0000 (10:57 +0000)]
[X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS
Ignoring freeze(undef) if it has multiple uses in LowerAVXCONCAT_VECTORS
causes the custom INSERT_SUBVECTOR for vector widening to be ignored.
Differential Revision: https://reviews.llvm.org/D144903
Balázs Kéri [Thu, 9 Mar 2023 10:13:52 +0000 (11:13 +0100)]
[clang][analyzer][NFC] Refactor code of StdLibraryFunctionsChecker.
The code was difficult to maintain (big internal class definitions
with long inline functions, other functions of the same class at
different location far away, irregular ordering of classes and
function definitions). It is now improved to some extent.
New functions are added to RangeConstraint to remove code repetition,
these are useful for planned new features too.
Comments are improved.
Reviewed By: Szelethus
Differential Revision: https://reviews.llvm.org/D143751
Haojian Wu [Thu, 9 Mar 2023 10:36:37 +0000 (11:36 +0100)]
Sven van Haastregt [Thu, 9 Mar 2023 10:20:04 +0000 (10:20 +0000)]
[OpenCL] Add builtin header TableGen emitter
Add an emitter to produce something similar to opencl-c.h from the
OpenCL builtin descriptions in OpenCLBuiltins.td
This only adds the emitter, without any direct use of it. This allows
opencl-c.h additions to be generated from the builtin descriptions by
manually invoking `clang-tblgen -gen-clang-opencl-builtin-header`.
Differential Revision: https://reviews.llvm.org/D104040
Petar Avramovic [Wed, 8 Mar 2023 16:06:29 +0000 (17:06 +0100)]
Fix SGPR + VGPR + offset Scratch offset folding
Values in SGPR and VGPR register are treated as unsigned by hardware.
When value in 32-bit SGPR or VGPR base can be negative calculate offset
using 32-bit add instructions, otherwise use
sgpr(unsigned) + vgpr(unsigned) + offset.
LoopStrengthReduce.cpp changes offsets to negative and in some
iterations value in SGPR or VGPR register could be negative.
Differential Revision: https://reviews.llvm.org/D144957
Petar Avramovic [Wed, 8 Mar 2023 15:42:14 +0000 (16:42 +0100)]
Fix VGPR + offset Scratch offset folding
Values in VGPR register are treated as unsigned by hardware.
When value in 32-bit VGPR base can be negative calculate offset using
32-bit add instruction, otherwise use vgpr base(unsigned) + offset.
Does not affect case where whole offset comes from VGPR register
(immediate offset is 0).
LoopStrengthReduce.cpp changes offsets to negative and in some
iterations value in VGPR register could be negative.
Differential Revision: https://reviews.llvm.org/D144956
Petar Avramovic [Wed, 8 Mar 2023 15:18:56 +0000 (16:18 +0100)]
Fix SGPR + offset Scratch offset folding
Values in SGPR register are treated as unsigned by hardware.
When value in 32-bit SGPR base can be negative calculate offset using
32-bit add instruction, otherwise use sgpr base(unsigned) + offset.
Does not affect case where whole offset comes from SGPR register
(immediate offset is 0).
LoopStrengthReduce.cpp changes offsets to negative and in some
iterations value in SGPR register could be negative.
Differential Revision: https://reviews.llvm.org/D144955
Yeting Kuo [Tue, 7 Mar 2023 12:06:19 +0000 (20:06 +0800)]
[IR][DAG][RISCV] Allow scalable vector ISD::STRICT_FP_EXTEND and RISC-V supports for vector ISD::STRICT_FP_EXTEND.
The patch mainly does two things. The first is allowing scalable vector
ISD::STRICT_FP_EXTEND. The second is making RISC-V customized lower
strict_fpextend to riscv_strict_fpextend_vl, the strict version of
riscv_fpextend_vl.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D145548
Chuanqi Xu [Thu, 9 Mar 2023 09:25:59 +0000 (17:25 +0800)]
[Modules] Remove unnecessary check when generating name lookup table in ASTWriter
Close https://github.com/llvm/llvm-project/issues/61065.
We will avoid writing the names from external AST naturally. But
currently its check is often false positive since we may have already
marked the declarations as external but
DeclContext::hasNeedToReconcileExternalVisibleStorage would be false
after reconciling.
Tested with libcxx's modular build.
This patch can improve 8% compilation time in an internal workloads.
Nikita Popov [Tue, 21 Feb 2023 11:00:11 +0000 (12:00 +0100)]
[IPSCCP] Don't require noundef for !range annotation
Since https://reviews.llvm.org/D141386 !range violations return
poison instead of causing immediate undefined behavior. As such,
it is fine for IPSCCP to infer !range even if the value might be
poison. (The value cannot be undef as this would promote undef to
poison, but this is already checked separately.)
This basically undoes the late change done to D83952, restoring
it to its original version (which is now valid).
Differential Revision: https://reviews.llvm.org/D144467
Nikita Popov [Thu, 9 Mar 2023 09:18:20 +0000 (10:18 +0100)]
[LangRef] Use new memory attribute syntax (NFC)
Use the memory() spelling in a few places that were using the
old syntax.
The documented attributes for llvm.type.checked.load don't match
the actual attributes, I've raised this here:
https://reviews.llvm.org/D21121#inline-1406792
Han Zhu [Thu, 9 Mar 2023 09:16:04 +0000 (01:16 -0800)]
[SROA] Fix bug where CandidateTys is appended while being iterated
Fix a crash when compiling Skia. See https://reviews.llvm.org/D143225#4180342
for more details
Marco Elver [Tue, 7 Mar 2023 19:00:12 +0000 (20:00 +0100)]
[SanitizerBinaryMetadata] Do not add to GPU code
SanitizerBinaryMetadata should only apply to to host code, and not GPU
code. Recently AMD GPU target code has experimental sanitizer support.
If we're compiling a mixed host/device source file, only add sanitizer
metadata to host code.
Differential Revision: https://reviews.llvm.org/D145519
David Spickett [Thu, 9 Mar 2023 09:14:13 +0000 (09:14 +0000)]
[llvm][TableGen][Jupyter] Correct notebook name in example command
Nikita Popov [Mon, 6 Mar 2023 15:53:35 +0000 (16:53 +0100)]
[llvm-c] Remove PassManagerBuilder APIs
The legacy PM is only supported for codegen, and PassManagerBuilder
is exclusively about the middle-end optimization pipeline. Drop it.
Differential Revision: https://reviews.llvm.org/D145387