David Green [Thu, 19 Dec 2019 10:40:17 +0000 (10:40 +0000)]
[Instcombine] Add select canonicalization tests. NFC
Hans Wennborg [Thu, 19 Dec 2019 12:23:13 +0000 (13:23 +0100)]
Re-land "Add an -fno-temp-file flag for compilation"
This time making sure to initialize FrontendOptions::UseTemporary.
Patch by Zachary Henkel!
Differential revision: https://reviews.llvm.org/D70615
Miloš Stojanović [Thu, 19 Dec 2019 11:52:15 +0000 (12:52 +0100)]
[llvm-exegesis] Fix pfm counter names for Haswell for older versions of libpfm
The inconsistency caused uops mode to fail on an older version of libpfm
since the dispatched_port was added as an alias for executed_port only
after v4.6.0 of libpfm.
Differential revision: https://reviews.llvm.org/D71665
Jay Foad [Thu, 19 Dec 2019 11:19:36 +0000 (11:19 +0000)]
Make more use of MachineInstr::mayLoadOrStore.
Victor Campos [Mon, 16 Dec 2019 14:22:15 +0000 (14:22 +0000)]
[ARM] Improve codegen of volatile load/store of i64
Summary:
Instead of generating two i32 instructions for each load or store of a volatile
i64 value (two LDRs or STRs), now emit LDRD/STRD.
These improvements cover architectures implementing ARMv5TE or Thumb-2.
Reviewers: dmgreen, efriedma, john.brawn
Reviewed By: efriedma
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D70072
Cullen Rhodes [Fri, 13 Dec 2019 16:32:35 +0000 (16:32 +0000)]
[AArch64][SVE] Implement pfirst and pnext intrinsics
Reviewers: sdesmalen, efriedma, dancgr, mgudim, cameron.mcinally
Reviewed By: cameron.mcinally
Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl,
llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71472
Cullen Rhodes [Thu, 12 Dec 2019 15:24:30 +0000 (15:24 +0000)]
[AArch64][SVE] Implement ptrue intrinsic
Reviewers: sdesmalen, eli.friedman, dancgr, mgudim, cameron.mcinally,
huntergr, efriedma
Reviewed By: sdesmalen
Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl,
llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71457
Kadir Cetinkaya [Tue, 17 Dec 2019 11:13:28 +0000 (12:13 +0100)]
[clangd] Improve documentation for auto and implicit specs
Summary:
Clangd didn't fill documentation for `auto` when it wasn't available in
index. Also it wasn't showing any documentations for implicit instantiations.
This patch ensures auto and normal decl case behaves in the same way and also
makes use of the explicit template specialization while fetching comments for
implicit instantiations.
Reviewers: ilya-biryukov
Subscribers: MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D71596
David Spickett [Thu, 19 Dec 2019 10:17:53 +0000 (10:17 +0000)]
[cmake] Include the llvm-locstats target when
utils and tools are not being built.
This was uncovered by: https://reviews.llvm.org/D71611
Which added llvm-locstats to the test dependencies.
Previously the build target was only added if you
were building tools. This meant that you couldn't
configure at all if you had LLVM_BUILD_TOOLS=OFF.
Differential revision: https://reviews.llvm.org/D71695
Dmitri Gribenko [Thu, 19 Dec 2019 10:24:51 +0000 (11:24 +0100)]
Revert "Support powerpc when builing without init_array."
This reverts commit
5789e83dedb97588ad75cca36d01ba6c5142d6d3. It broke
the build on aarch64.
Martin Storsjö [Wed, 18 Dec 2019 08:41:41 +0000 (10:41 +0200)]
[clang] Fix compilation with GCC < 8 for MinGW
GCC 7 and earlier, when targeting MinGW, seems to have a bug in
layout/size of bitfield structs if they contain a nested enum,
making the size of the struct 8 bytes, while we have a static assert
requiring it to be 4 bytes or less.
While this clearly is a GCC bug, the workaround (moving the enum out
of the bitfield) also is very nonintrusive and matches other existing
enums there.
Differential Revision: https://reviews.llvm.org/D71650
Martin Storsjö [Wed, 18 Dec 2019 21:58:51 +0000 (23:58 +0200)]
[LLD] [COFF] Fix reporting duplicate errors for absolute symbols
Previously this caused crashes in the reportDuplicate method.
A DefinedAbsolute doesn't have any InputFile attached to it, so we
can't report the file for the original symbol.
We could add an InputFile argument to SymbolTable::addAbsolute
only for the sake of error reporting, but even then it'd be assymetrical,
only pointing out the file containing the new conflicting definition,
not the original one.
Differential Revision: https://reviews.llvm.org/D71679
Raphael Isemann [Thu, 19 Dec 2019 09:34:28 +0000 (10:34 +0100)]
[lldb][NFC] Change if statements in ClangASTImporter to follow LLVM code style
czhengsz [Thu, 19 Dec 2019 09:11:45 +0000 (04:11 -0500)]
[SCEV] NFC - add testcase for get accurate range for AddExpr
Georgii Rymar [Tue, 17 Dec 2019 10:54:29 +0000 (13:54 +0300)]
[llvm-readobj][test] - Improve dyn-symbols.test.
This removes the precompiled binary used, simplifies
the first test case, adds comments and llvm-readelf tool
invocations.
It also adds a test case for checking versioning symbols.
Differential revision: https://reviews.llvm.org/D71595
Georgii Rymar [Wed, 18 Dec 2019 09:44:54 +0000 (12:44 +0300)]
[llvm-readobj][test] - Refactor mips-st-other.test
This removes 2 precompiled binaries, adds testing
for STO_* flags missing, refines and renames the test.
Differential revision: https://reviews.llvm.org/D71651
Georgii Rymar [Tue, 17 Dec 2019 12:53:36 +0000 (15:53 +0300)]
[llvm-readobj][llvm-objdump] - Cleanup testing of dynamic tags dumping.
We have the `elf-dynamic-tags-machine-specific.yaml` input shared
between the llvm-readobj and llvm-objdump test.
It looks strange, because tools usually does not share inputs.
Also there are following problems related:
1) `elf-dynamic-tags-machine-specific.yaml` input contains excessive YAML parts.
2) objdump's test case never test AARCH64 tags.
3) There are unknown tags in the `elf-dynamic-tags-machine-specific.yaml` and
`dynamic-tags-machine-specific.test`, though we already testing unknown tags
in `\llvm-readobj\ELF\dynamic-tags.test` and `llvm-objdump\elf-dynamic-section.test` tests.
This patch removes the shared input and refines the test cases to resolve
issues mentioned.
Differential revision: https://reviews.llvm.org/D71602
Stanislav Mekhanoshin [Wed, 18 Dec 2019 21:29:21 +0000 (13:29 -0800)]
[AMDGPU] Implemented fma cost analysis
Differential Revision: https://reviews.llvm.org/D71676
Liu, Chen3 [Thu, 19 Dec 2019 06:43:45 +0000 (14:43 +0800)]
Enable STRICT_FP_TO_SINT/UINT on X86 backend
This patch is mainly for custom lowering the vector operation.
Differential Revision: https://reviews.llvm.org/D71592
David Blaikie [Thu, 19 Dec 2019 03:00:40 +0000 (19:00 -0800)]
Kamil Rytarowski [Thu, 19 Dec 2019 02:21:46 +0000 (03:21 +0100)]
[compiler-rt] Enable SANITIZER_CAN_USE_PREINIT_ARRAY on NetBSD
.preinit_array is supported since 9.0.
czhengsz [Thu, 19 Dec 2019 02:03:54 +0000 (21:03 -0500)]
[PowerPC] make lwa as a valid ds candidate in ppcloopinstrformprep pass
Fix a FIXME in ppcloopinstrformprep pass.
Reviewed by: nemanjai
Differential Revision: https://reviews.llvm.org/D71346
JonChesterfield [Thu, 19 Dec 2019 02:00:26 +0000 (02:00 +0000)]
[libomptarget][nvptx] Fix build, second symbol reordering
Artem Dergachev [Thu, 19 Dec 2019 01:59:16 +0000 (17:59 -0800)]
[analysis] Re-discard type sugar when casting values retrieved from the Store.
Canonicalization was accidentally omitted in
6d3f43ec.
Jon Chesterfield [Thu, 19 Dec 2019 01:50:06 +0000 (01:50 +0000)]
[libomptarget][nvptx] Fix build, symbol ordering in target_impl.h
David Blaikie [Thu, 19 Dec 2019 01:10:24 +0000 (17:10 -0800)]
DebugInfo: Include DW_AT_base_addr even in gmlt with no inline functions
Since the address pool doesn't get populated in this case (due to the
lack of inlining, no child DIEs are added to the CU - so no addresses
are needed for the DIEs themselves) until the range list is emitted - at
the time the attributes are added to the CU, the address pool is empty.
So check whether the address pool will be used for the range lists & add
an addr_base if that's the case.
Jordan Rupprecht [Thu, 19 Dec 2019 00:44:53 +0000 (16:44 -0800)]
[clang] Remove -Wexperimental-float-control.
Summary: Per D62731, the behavior of clang with `-frounding-math` is no worse than when the rounding flag was completely ignored, so remove this unnecessary warning.
Reviewers: mibintc, chandlerc, echristo, rjmccall, kpn, erichkeane, rsmith, andrew.w.kaylor
Reviewed By: mibintc
Subscribers: merge_guards_bot, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D71671
Jonas Devlieghere [Thu, 19 Dec 2019 00:42:07 +0000 (16:42 -0800)]
[lldb/Scripting] Simplify code by removing the #if.
The "none" script interpreter does not depend on Python so it doesn't
make sense to have it withing the if-block. The only goal seems to be to
have a slightly different error for when there's no script interpreter,
but as per the comment this doesn't make sense for more than one
scripting language. I think the existing error is perfectly clear, so I
just removed this altogether.
Francis Visoiu Mistrih [Wed, 18 Dec 2019 23:55:30 +0000 (15:55 -0800)]
[Remarks][Driver] Place temporary remark files next to temporary object files
On Darwin, when used for generating a linked binary from a source file
(through an intermediate object file), the driver will invoke `cc1` to
generate a temporary object file. The temporary remark file will now be
emitted next to the object file, which will then be picked up by
`dsymutil` and emitted in the .dSYM bundle.
This is available for all formats except YAML since by default, YAML
doesn't need a section and the remark file will be lost.
Francis Visoiu Mistrih [Wed, 18 Dec 2019 23:27:21 +0000 (15:27 -0800)]
[Docs] Fix indentation in remarks section
David Blaikie [Thu, 19 Dec 2019 00:19:17 +0000 (16:19 -0800)]
DebugInfo: Don't use implicit zero addr_base
(found when LLVM fails to emit addr_base for gmlt+DWARFv5)
David Blaikie [Mon, 16 Dec 2019 23:19:25 +0000 (15:19 -0800)]
Reapply "NFC: DebugInfo: Refactor RangeSpanList to be a struct, like DebugLocStream::List"
Move these data structures closer together so their emission code can
eventually share more of its implementation.
Was an egregious bug (completely untested, evidently) where I hadn't
inverted a DWARFv5 test as needed, so it was doing the exact opposite of
what was required & thus tried to emit a DWARFv5 range list header in
DWARFv4.
Reapply
8e04896288d22ed8bef7ac367923374f96b753d6 which was
reverted in
a8154e5e0c83d2f0f65f3b4fb1a0bc68785bd975.
Kamil Rytarowski [Thu, 19 Dec 2019 00:15:03 +0000 (01:15 +0100)]
[compiler-rt] [fuzzer] Enable LSan in libFuzzer tests on NetBSD
Jason Molenda [Thu, 19 Dec 2019 00:11:06 +0000 (16:11 -0800)]
make err msg in MachSymtabSectionInfo::GetSection clear about the file
This error message didn't specify which file was malformed, so
there's some hunting-around required if it comes up. We have the
filename; include it in the error message.
Kamil Rytarowski [Thu, 19 Dec 2019 00:08:26 +0000 (01:08 +0100)]
[compiler-rt] [fuzzer] Include stdarg.h for va_list
Fixes build on NetBSD after
139e216e6610091b7ee3c30bc11114f5d73cbd3e.
JonChesterfield [Thu, 19 Dec 2019 00:11:21 +0000 (00:11 +0000)]
[libomptarget][amdgcn] Correct return type of extern __clock64 to unsigned
Jonas Devlieghere [Wed, 18 Dec 2019 23:41:07 +0000 (15:41 -0800)]
[lldb/test] Move script interpreter tests
Create a new test for lldb launched without a script interpreter and
move it under a new `ScriptInterpreter` directory. Also move
crashlog.test there for consistency.
Thomas Lively [Wed, 18 Dec 2019 05:53:46 +0000 (21:53 -0800)]
[WebAssembly] Add avgr_u intrinsics and require nuw in patterns
Summary:
The vector pattern `(a + b + 1) / 2` was previously selected to an
avgr_u instruction regardless of nuw flags, but this is incorrect in
the case where either addition may have an unsigned wrap. This CL
changes the existing pattern to require both adds to have nuw flags
and adds builtin functions and intrinsics for the avgr_u instructions
because the corrected pattern is not representable in C.
Reviewers: aheejin
Subscribers: dschuff, sbc100, jgravelle-google, hiraditya, sunfish, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71648
Yuanfang Chen [Wed, 18 Dec 2019 23:19:01 +0000 (15:19 -0800)]
[Docs] Fix a typo
Lang Hames [Wed, 18 Dec 2019 23:13:35 +0000 (15:13 -0800)]
Revert "[Orc][LLJIT] Use JITLink even if a custom JITTargetMachineBuilder is supplied."
This reverts commit
298e183e813c884dd22816383405bae3ef9ef278.
This commit caused some build failures -- reverting while I investigate.
Vedant Kumar [Wed, 18 Dec 2019 22:52:51 +0000 (14:52 -0800)]
Revert concepts changes from D41910
These changes caused LibcxxVariantDataFormatterTestCase in lldb to fail
with an assert:
Assertion failed: (Idx < size() && "Out-of-bounds Bit access."),
function operator[], file
/Users/buildslave/jenkins/workspace/lldb-cmake/llvm-project/llvm/include/llvm/ADT/SmallBitVector.h,
line 452.
In:
7 clang-10 0x00000001094b79d9 isAtLeastAsSpecializedAs(clang::Sema&, clang::SourceLocation, clang::FunctionTemplateDecl*, clang::FunctionTemplateDecl*, clang::TemplatePartialOrderingContext, unsigned int) + 1865
8 clang-10 0x00000001094b7111 clang::Sema::getMoreSpecializedTemplate(clang::FunctionTemplateDecl*, clang::FunctionTemplateDecl*, clang::SourceLocation, clang::TemplatePartialOrderingContext, unsigned int, unsigned int) + 97
9 clang-10 0x000000010939bf88 clang::isBetterOverloadCandidate(clang::Sema&, clang::OverloadCandidate const&, clang::OverloadCandidate const&, clang::SourceLocation, clang::OverloadCandidateSet::CandidateSetKind) + 1128
Revert "[Concepts] Fix incorrect move out of temporary in D41910"
This reverts commit
11d5fa6e87e3584f72056ecc2b17f88c58323dde.
Revert "[Concepts] Fix crash in D41910"
This reverts commit
12038be20ee6a903cdbd3fddce65535ef683e31d.
Revert "[Concepts] Constrained partial specializations and function overloads."
This reverts commit
fc0731b98a67c793862288f8ae334322666214dc.
Craig Topper [Wed, 18 Dec 2019 22:24:20 +0000 (14:24 -0800)]
[X86] Add a simple hack to IsProfitableToFold to prevent vselect+strict fp operations from being folded into masked instructions.
We really need to update the isel patterns to prevent this, but
that requires some tablegen de-tangling. So this hack will work
for correctness in the short term.
Craig Topper [Wed, 18 Dec 2019 22:19:55 +0000 (14:19 -0800)]
[X86] Add test case to show incorrect formation of masked add from vselect+strict_fadd.
The masked operation will suppress exceptions on masked out lanes.
But that's not what vselect+strict_fadd means.
Stephen Kelly [Sat, 4 May 2019 15:51:58 +0000 (16:51 +0100)]
Add method to ignore invisible AST nodes
Reviewers: aaron.ballman
Subscribers: mgorny, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D70613
Stephen Kelly [Wed, 18 Dec 2019 22:12:10 +0000 (22:12 +0000)]
Trim trailing whitespace
Stephen Kelly [Wed, 18 Dec 2019 21:50:50 +0000 (21:50 +0000)]
Output names in the AST in tests
Francis Visoiu Mistrih [Wed, 18 Dec 2019 19:13:46 +0000 (11:13 -0800)]
[Remarks][Driver] Run dsymutil when remarks are enabled
When clang is invoked with a source file without -c or -S, it creates a
cc1 job, a linker job and if debug info is requested, a dsymutil job. In
case of remarks, we should also create a dsymutil job to avoid losing
the remarks that will be generated in a tempdir that gets removed.
Differential Revision: https://reviews.llvm.org/D71675
Francis Visoiu Mistrih [Wed, 18 Dec 2019 21:16:09 +0000 (13:16 -0800)]
[Remarks][Driver][NFC] Make shouldEmitRemarks more available in the Driver
Move the function to Driver.h so that it can be re-used in other places.
Artem Dergachev [Wed, 18 Dec 2019 21:19:44 +0000 (13:19 -0800)]
[analyzer] Teach MismatchedDealloc about initWithBytesNoCopy with deallocator.
MallocChecker warns when memory is passed into -[NSData initWithBytesNoCopy]
but isn't allocated by malloc(), because it will be deallocated by free().
However, initWithBytesNoCopy has an overload that takes an arbitrary block
for deallocating the object. If such overload is used, it is no longer
necessary to make sure that the memory is allocated by malloc().
Jason Molenda [Wed, 18 Dec 2019 22:18:22 +0000 (14:18 -0800)]
Add prototype for a function we call.
Lang Hames [Mon, 16 Dec 2019 10:45:51 +0000 (02:45 -0800)]
[Orc][LLJIT] Use JITLink even if a custom JITTargetMachineBuilder is supplied.
LLJITBuilder will now use JITLink on supported platforms even if a custom
JITTargetMachineBuilder is supplied, provided that neither the code model,
nor the relocation model, nor the ObjectLinkingLayerCreator is set.
Richard Smith [Wed, 18 Dec 2019 22:01:40 +0000 (14:01 -0800)]
Refactor CompareReferenceRelationship and its callers in preparation for
implementing the resolution of CWG2352.
No functionality change, except that we now convert the referent of a
reference binding to the underlying type of the reference in more cases;
we used to happen to preserve the type sugar from the referent if the
only type change was in the cv-qualifiers.
This exposed a bug in how we generate code for trivial assignment
operators: if the type sugar (particularly the may_alias attribute)
got lost during reference binding, we'd use the "wrong" TBAA information
for the load during the assignment.
Akira Hatanaka [Wed, 18 Dec 2019 21:54:30 +0000 (13:54 -0800)]
Use hasOffsetApplied to initialize member HasOffsetApplied
This is NFC since none of the constructor calls in trunk pass
hasOffsetApplied=true.
Reid Kleckner [Sat, 7 Dec 2019 00:10:28 +0000 (16:10 -0800)]
Move TypeSourceInfo to Type.h
TypeSourceInfo is a thin wrapper around TypeLocs. Notionally, the best
place for it to live would be TypeLoc.h, but Decl.h requires it to be
complete, so it needs to be lower in the dependency graph. Type.h seems
like the next best place.
By itself, this change has no impact on build time, because it doesn't
remove a single dependency edge from a .cpp file to a .h file, but it is
an incremental step towards making the AST headers less interdependent.
Reviewers: rsmith
Differential Revision: https://reviews.llvm.org/D71427
Sterling Augustine [Wed, 18 Dec 2019 19:15:32 +0000 (11:15 -0800)]
Support powerpc when builing without init_array.
Summary: Also add an error case when targetting an unimplement architecture.
Subscribers: nemanjai, jsji, shchenz, #sanitizers, llvm-commits
Tags: #sanitizers, #llvm
Differential Revision: https://reviews.llvm.org/D71670
Michał Górny [Wed, 18 Dec 2019 21:10:20 +0000 (22:10 +0100)]
[clang-tools-extra] Fix linking dylib for LLVMFrontendOpenMP
Use LLVM_LINK_COMPONENTS to link the FrontendOpenMP library
instead of passing it explicitly to LINK_LIBS. This fixes duplicating
the library when clang-tidy is linked to LLVM dylib.
Differential Revision: https://reviews.llvm.org/D71674
Amy Huang [Wed, 18 Dec 2019 21:15:13 +0000 (13:15 -0800)]
Change triple in test case to not include triples the test shouldn't
pass.
This is fixes changes from
a85f5efd9597d0036f5c347b362cb873bdf51f16.
Mehdi Amini [Wed, 18 Dec 2019 21:11:24 +0000 (16:11 -0500)]
[gn build] Update console print to match the commit message (NFC)
Piotr Sobczak [Wed, 18 Dec 2019 21:01:04 +0000 (22:01 +0100)]
Revert "[InstCombine][AMDGPU] Trim more components of *buffer_load"
Revert D70315, as it breaks gfx8 for some reason.
This reverts commit
65f94b33808d7d69539961a6f5a2168f0a1eef41.
Mehdi Amini [Wed, 18 Dec 2019 20:55:27 +0000 (15:55 -0500)]
Tweak commit message for auto update of gn build script
Use `[tag]` as a prefix and "port" instead of "merge" to
avoid confusing with "git merge".
Jonas Devlieghere [Wed, 18 Dec 2019 20:49:46 +0000 (12:49 -0800)]
[lldb/MachO] "Fix" intentional out-of-bounds error (NFC)
Remove the hack that populates the cpsr register in the gpr struct by
writing past the end of the array. This was tripping up ASan.
Patch by: Reva Cuthbertson
Ulrich Weigand [Wed, 18 Dec 2019 20:33:10 +0000 (21:33 +0100)]
[FPEnv] Strict versions of llvm.minimum/llvm.maximum
Add new intrinsics
llvm.experimental.constrained.minimum
llvm.experimental.constrained.maximum
as strict versions of llvm.minimum and llvm.maximum.
Includes SystemZ back-end support.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D71624
Steven Wu [Wed, 18 Dec 2019 20:22:21 +0000 (12:22 -0800)]
[libunwind] Fix evaluating DWARF operation DW_OP_pick
reg is unsigned type and used here for getting array element from the end by
negating it. negation of unsigned can result in large number and array access
with that index will result in segmentation fault.
Fixes: https://bugs.llvm.org/show_bug.cgi?id=43872
Patched by: kamlesh kumar
Differential Revision: https://reviews.llvm.org/D69893
Kamil Rytarowski [Wed, 18 Dec 2019 20:03:17 +0000 (21:03 +0100)]
Correct version check 9.9.26 -> 9.99.26
Artem Dergachev [Wed, 18 Dec 2019 20:04:18 +0000 (12:04 -0800)]
[analyzer] NonnullGlobalConstants: Add support for kCFNull.
It's a singleton in CoreFoundation that always contains a non-null CFNullRef.
Jon Chesterfield [Wed, 18 Dec 2019 20:06:16 +0000 (20:06 +0000)]
[libomptarget][nfc] Introduce atomic wrapper function
Summary:
[libomptarget][nfc] Introduce atomic wrapper function
Wraps atomic functions in a template prefixed __kmpc_atomic that
dispatches to cuda or hip atomic functions. Intended to be easily extended
to dispatch to OpenCL or C++ atomics for a third target.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: Anastasia, jvesely, mgrang, dexonsmith, llvm-commits, mgorny, jfb, openmp-commits
Tags: #openmp, #llvm
Differential Revision: https://reviews.llvm.org/D71404
Kit Barton [Mon, 16 Dec 2019 21:37:30 +0000 (16:37 -0500)]
[LoopFusion] Use the LoopInfo::isRotatedForm method (NFC).
Loop fusion previously had a method to check whether a loop was in rotated form. This method has
been moved into the LoopInfo class. This patch removes the old isRotated method from loop fusion,
in favour of the new one in LoopInfo.
Jakub Kuderski [Wed, 18 Dec 2019 19:55:41 +0000 (14:55 -0500)]
[InstCombine] Insert instructions before adding them to worklist
Summary:
This patch adds instructions to the InstCombine worklist after they are properly inserted. This way we don't get `<badref>`s printed when logging added instructions.
It also adds a check in `Worklist::Add` that ensures that all added instructions have parents.
Simple test case that illustrates the difference when run with `--debug-only=instcombine`:
```
define i32 @test35(i32 %a, i32 %b) {
%1 = or i32 %a, 1135
%2 = or i32 %1, %b
ret i32 %2
}
```
Before this patch:
```
INSTCOMBINE ITERATION #1 on test35
IC: ADDING: 3 instrs to worklist
IC: Visiting: %1 = or i32 %a, 1135
IC: Visiting: %2 = or i32 %1, %b
IC: ADD: %2 = or i32 %a, %b
IC: Old = %3 = or i32 %1, %b
New = <badref> = or i32 %2, 1135
IC: ADD: <badref> = or i32 %2, 1135
...
```
With this patch:
```
INSTCOMBINE ITERATION #1 on test35
IC: ADDING: 3 instrs to worklist
IC: Visiting: %1 = or i32 %a, 1135
IC: Visiting: %2 = or i32 %1, %b
IC: ADD: %2 = or i32 %a, %b
IC: Old = %3 = or i32 %1, %b
New = <badref> = or i32 %2, 1135
IC: ADD: %3 = or i32 %2, 1135
...
```
Reviewers: fhahn, davide, spatel, foad, grosser, nikic
Reviewed By: nikic
Subscribers: nikic, lebedev.ri, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71093
Saar Raz [Wed, 18 Dec 2019 19:43:53 +0000 (21:43 +0200)]
[Concepts] Fix incorrect move out of temporary in D41910
Moves out of temporaries caused warnings that failed builds.
JonChesterfield [Wed, 18 Dec 2019 19:39:34 +0000 (19:39 +0000)]
[libomptarget][nfc] Extract function from data_sharing, move to common
Summary:
[libomptarget][nfc] Extract function from data_sharing, move to common
Finding the first active thread in the warp is different on nvptx and amdgcn,
mostly due to warp size and the desire for efficiency.
Reviewers: ABataev, jdoerfert, grokos
Reviewed By: jdoerfert
Subscribers: jvesely, mgorny, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D71643
Adrian McCarthy [Wed, 4 Dec 2019 21:11:02 +0000 (13:11 -0800)]
[NFC] Update FIXME for one VFS test
The VFS/subframework-symlink.m test is still XFAIL on Windows, but for
a different reason than those fixed in PR43272, so I've updated the
PR number.
Adrian McCarthy [Mon, 25 Nov 2019 23:57:21 +0000 (15:57 -0800)]
Fix more VFS tests on Windows
Since VFS paths can be in either Posix or Windows style, we have to use
a more flexible definition of "absolute" path.
The key here is that FileSystem::makeAbsolute is now virtual, and the
RedirectingFileSystem override checks for either concept of absolute
before trying to make the path absolute by combining it with the current
directory.
Differential Revision: https://reviews.llvm.org/D70701
Saar Raz [Wed, 18 Dec 2019 19:31:33 +0000 (21:31 +0200)]
[Concepts] Fix crash in D41910
Differential Revision: https://reviews.llvm.org/D41910
Danilo Carvalho Grael [Wed, 18 Dec 2019 19:01:38 +0000 (14:01 -0500)]
Revert "[AArch64][SVE] Replace integer immediate intrinsics with splat vector variant"
This reverts commit
830e08b98bcb427136443093c282b25328137cf0 and
eb1857ce0da481caf82271e6d0c9fc745dfab26f.
This commit leads to an unexpected failure on test/CodeGen/AArch64/sve-gather-scatter-dag-combine.ll.
The review will need more changes before its re-commited.
Jonas Paulsson [Wed, 18 Dec 2019 18:12:41 +0000 (10:12 -0800)]
[Clang FE, SystemZ] Don't add "true" value for the "mnop-mcount" attribute.
Let the "mnop-mcount" function attribute simply be present or non-present.
Update SystemZ backend as well to use hasFnAttribute() instead.
Review: Ulrich Weigand
https://reviews.llvm.org/D71669
Saar Raz [Wed, 18 Dec 2019 18:59:01 +0000 (20:59 +0200)]
[Concepts] Constrained partial specializations and function overloads.
Added support for constraint satisfaction checking and partial ordering of constraints in constrained partial specialization and function template overloads.
Phabricator: D41910
Jakub Kuderski [Wed, 18 Dec 2019 18:48:54 +0000 (13:48 -0500)]
[InstCombine] Allow to limit the max number of iterations
Summary:
This patch teaches InstCombine to accept a new parameter: maximum number of iterations over functions.
InstCombine tries to simplify instructions by iterating over the whole function until the function stops changing. As a consequence, the last iteration before reaching a fixpoint visits all instructions in the worklist and never performs any rewrites.
Bounding the number of iterations can have 2 benefits:
* In case the users of the pass can make a good guess about the number of required iterations, we can save the time normally spent on the last iteration that doesn't change anything.
* When the wants to use InstCombine as a cleanup pass, it may be enough to run just a few iterations and stop even before reaching a fixpoint. This can be also useful for implementing a lightweight pass pipeline (think `-O1`).
This patch does not change the behavior of opt or Clang -- limiting the number of iterations is entirely opt-in.
Reviewers: fhahn, davide, spatel, foad, nlopes, grosser, lebedev.ri, nikic, xbolva00
Reviewed By: spatel
Subscribers: craig.topper, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71145
Richard Smith [Wed, 18 Dec 2019 18:44:29 +0000 (10:44 -0800)]
llvm-cxxmap: fix support for remapping non-mangled names.
Remappings involving extern "C" names were already supported in the
context of <local-name>s, but this support didn't work for remapping the
complete mangling itself. (Eg, we would remap X<foo> but not foo itself,
if foo is an extern "C" function.)
Craig Topper [Wed, 18 Dec 2019 18:41:34 +0000 (10:41 -0800)]
[SelectionDAGBuilder] Use getConstant instead of getTargetConstant to build the offset for struct types in getUniformBase.
getTargetConstant prevents any optimizations from operating on the
value and basically says its already been iseled. But since we
want the index to be in a register, this isn't true.
Prior to this we were generating a vbroadcast with an immediate
argument which is illegal and was flagged by the expensive checks
bot.
Stefan Pintilie [Wed, 18 Dec 2019 18:19:04 +0000 (12:19 -0600)]
[PowerPC][NFC] Refactor splat of constant to vector.
Refactor the splatting of a constant to a vector so that common code is used
both for Power9 and Power8.
Patch by: Anil Mahmud
Differential Revision: https://reviews.llvm.org/D71481
Amy Huang [Thu, 24 Oct 2019 23:34:25 +0000 (16:34 -0700)]
Add support for the MS qualifiers __ptr32, __ptr64, __sptr, __uptr.
Summary:
This adds parsing of the qualifiers __ptr32, __ptr64, __sptr, and __uptr and
lowers them to the corresponding address space pointer for 32-bit and 64-bit pointers.
(32/64-bit pointers added in https://reviews.llvm.org/D69639)
A large part of this patch is making these pointers ignore the address space
when doing things like overloading and casting.
https://bugs.llvm.org/show_bug.cgi?id=42359
Reviewers: rnk, rsmith
Subscribers: jholewinski, jvesely, nhaehnle, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D71039
Danilo Carvalho Grael [Wed, 18 Dec 2019 18:44:25 +0000 (13:44 -0500)]
[AArch64][SVE] Fix gather scatter dag combine test.
Danilo Carvalho Grael [Tue, 17 Dec 2019 15:42:52 +0000 (10:42 -0500)]
[AArch64][SVE] Replace integer immediate intrinsics with splat vector variant
Summary: Replace the integer immediate intrisics with splat vector variants so they can be applied as optimizations for the C/C++ intrinsics.
Reviewers: sdesmalen, huntergr, rengolin, efriedma, c-rhodes, mgudim, kmclaughlin
Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl, llvm-commits, amehsan
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71614
Michael Trent [Wed, 11 Dec 2019 18:42:37 +0000 (10:42 -0800)]
[ MC ] Match labels to existing fragments even when switching sections.
(This commit restores the original branch (
4272372c571) and applies an
additional change dropped from the original in a bad merge. This change
should address the previous bot failures. Both changes reviewed by pete.)
Summary:
This commit builds upon Derek Schuff's 2014 commit for attaching labels to
existing fragments ( Diff Revision: http://reviews.llvm.org/D5915 )
When temporary labels appear ahead of a fragment, MCObjectStreamer will
track the temporary label symbol in a "Pending Labels" list. Labels are
associated with fragments when a real fragment arrives; otherwise, an empty
data fragment will be created if the streamer's section changes or if the
stream finishes.
This commit moves the "Pending Labels" list into each MCStream, so that
this label-fragment matching process is resilient to section changes. If
the streamer emits a label in a new section, switches to another section to
do other work, then switches back to the first section and emits a
fragment, that initial label will be associated with this new fragment.
Labels will only receive empty data fragments in the case where no other
fragment exists for that section.
The downstream effects of this can be seen in Mach-O relocations. The
previous approach could produce local section relocations and external
symbol relocations for the same data in an object file, and this mix of
relocation types resulted in problems in the ld64 Mach-O linker. This
commit ensures relocations triggered by temporary labels are consistent.
Reviewers: pete, ab, dschuff
Reviewed By: pete, dschuff
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71368
Sanjay Patel [Wed, 18 Dec 2019 16:55:26 +0000 (11:55 -0500)]
[AArch64] add tests for bitcasted DUPLANE; NFC
See D63815 for context/motivation.
Mitch Phillips [Wed, 18 Dec 2019 17:05:09 +0000 (09:05 -0800)]
Revert "Add an -fno-temp-file flag for compilation"
This reverts commit
d129aa1d5369781deff6c6b854cb612e160d3fb2.
This broke the MSan buildbots. More information available in the
original PR: https://reviews.llvm.org/D70615
Mitch Phillips [Wed, 18 Dec 2019 16:47:31 +0000 (08:47 -0800)]
[NFC] Fix sanitizer lint problem that's breaking Android bot.
Fixes the lint issue introduced by
e14ff22694c28b5e731502bfc30192b3208d4d8d
Raphael Isemann [Wed, 18 Dec 2019 16:42:58 +0000 (17:42 +0100)]
Fix modules build by adding missing includes to LTO/Config.h
stozer [Wed, 18 Dec 2019 14:28:24 +0000 (14:28 +0000)]
Reapply: [DebugInfo] Correctly handle salvaged casts and split fragments at ISel
This reverts commit
1f3dd83cc1f2b8f72b9d59e2b4221b12fb7f9a95, reapplying
commit
bb1b0bc4e57428ce364d3d6c075ff03cb8973462.
The original commit failed on some builds seemingly due to the use of a
bracketed constructor with an std::array, i.e. `std::array<> arr({...})`.
Guillaume Chatelet [Wed, 18 Dec 2019 11:08:38 +0000 (12:08 +0100)]
[llvm-exegesis][NFC] internal changes
Summary:
BitVectors are now cached to lower memory utilization.
Instructions have reference semantics.
Reviewers: courbet
Subscribers: sdardis, tschuett, jrtc27, atanasyan, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71653
Michał Górny [Wed, 18 Dec 2019 16:20:40 +0000 (17:20 +0100)]
[compiler-rt] [test] Disable ASLR on UBSAN+[AMT]SAN tests on NetBSD
Michał Górny [Wed, 18 Dec 2019 16:07:34 +0000 (17:07 +0100)]
[compiler-rt] [test] Disable ASLR on LSAN+ASAN tests on NetBSD
Mircea Trofin [Wed, 18 Dec 2019 15:56:47 +0000 (07:56 -0800)]
[NFC][InlineCost] Run clang-format on InlineCost.cpp
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: Jim, eraman, hiraditya, haicheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D71646
Sanjay Patel [Wed, 18 Dec 2019 16:00:38 +0000 (11:00 -0500)]
[AArch64] update test checks; NFC
The common prefix reduces a bunch of replication; not sure why it
didn't happen before.
LLVM GN Syncbot [Wed, 18 Dec 2019 16:00:24 +0000 (16:00 +0000)]
gn build: Merge
9883d7edc65
Whitney Tsang [Wed, 18 Dec 2019 15:57:50 +0000 (15:57 +0000)]
[LoopUtils] Updated deleteDeadLoop() to handle loop nest.
Reviewer: kariddi, sanjoy, reames, Meinersbur, bmahjour, etiotto,
kbarton
Reviewed By: Meinersbur
Subscribers: mgorny, hiraditya, llvm-commits
Tag: LLVM
Differential Revision: https://reviews.llvm.org/D70939
Michał Górny [Wed, 18 Dec 2019 15:51:31 +0000 (16:51 +0100)]
[compiler-rt] [test] Disable ASLR for fuzzer tests on NetBSD
Michał Górny [Tue, 17 Dec 2019 17:09:18 +0000 (18:09 +0100)]
[clang] [cmake] Fix gen_ast_dump_json_test.py binary dir
Use correct directory to put gen_ast_dump_json_test.py instead of
LLVM_BINARY_DIR. This fixes attempting to write to /usr when building
clang standalone.
Differential Revision: https://reviews.llvm.org/D71615
evgeny [Wed, 18 Dec 2019 15:33:15 +0000 (18:33 +0300)]
[ThinLTO] Show preserved symbols in DOT files
Differential revision: https://reviews.llvm.org/D71608
Daniel Sanders [Wed, 18 Dec 2019 14:56:23 +0000 (14:56 +0000)]
[globalisel] Small correction to the fixup in
0a0813962d6. The two ID's are not the same size
The rule ID is a uint64_t as it has global scope but the UID is scoped to
the rule and expected to be much smaller so it's only unsigned.