Simon Pilgrim [Sun, 4 Aug 2019 12:48:19 +0000 (12:48 +0000)]
Fix signed/unsigned comparison warning. NFC.
llvm-svn: 367783
Simon Pilgrim [Sun, 4 Aug 2019 12:24:40 +0000 (12:24 +0000)]
[X86] SimplifyMultipleUseDemandedBits - Add target shuffle support
llvm-svn: 367782
Benjamin Kramer [Sun, 4 Aug 2019 11:08:32 +0000 (11:08 +0000)]
[ORC] Remove a layer of indirection when locking the mutex. NFCI.
llvm-svn: 367781
David Green [Sun, 4 Aug 2019 10:18:15 +0000 (10:18 +0000)]
[ARM] MVE big endian bitcasts
This adds big endian MVE patterns for bitcasts. They are defined in llvm as
being the same as a store of the existing type and the load into the new. This
means that they have to become a VREV between the two types, working in the
same way that NEON works in big-endian. This also adds some example tests for
bigendian, showing where code is and isn't different.
The main difference, especially from a testing perspective is that vectors are
passed as v2f64, and so are VREV into and out of call arguments, and the
parameters are passed in a v2f64 format. Same happens for inline assembly where
the register class is used, so it is VREV to a v16i8.
So some of this is probably not correct yet, but it is (mostly) self-consistent
and seems to be consistent with how llvm treats vectors. The rest we can
hopefully fix later. More details about big endian neon can be found in
https://llvm.org/docs/BigEndianNEON.html.
Differential Revision: https://reviews.llvm.org/D65581
llvm-svn: 367780
Serge Pavlov [Sun, 4 Aug 2019 10:08:51 +0000 (10:08 +0000)]
[Parser] Emit descriptive diagnostic for misplaced pragma
If a class or struct or union declaration contains a pragma that
is not valid in this context, compiler issues generic error like
"expected member name or ';' after declaration specifiers". With this
change the error tells that this pragma cannot appear in this declaration.
Differential Revision: https://reviews.llvm.org/D64932
llvm-svn: 367779
Fangrui Song [Sun, 4 Aug 2019 09:05:21 +0000 (09:05 +0000)]
[ELF][test] Delete redundant version-script-*.s tests
Delete version-script-missing.s: it is covered by version-script-noundef.s
Delete version-script-anonymous-local.s: it is covered by version-script-{glob,weak}.s etc
Delete version-script-no-warn{,2}.s: add --fatal-warnings to some version-script.s commands instead
llvm-svn: 367778
Eric Fiselier [Sun, 4 Aug 2019 07:13:43 +0000 (07:13 +0000)]
Revert "Suppress -Wctad-maybe-unsupported on types w/o deduction guides."
Some modules builds are issuing buggy diagnostics. The cause of which is
TBD.
This reverts commit r@367770.
llvm-svn: 367777
Michael Pozulp [Sun, 4 Aug 2019 06:04:00 +0000 (06:04 +0000)]
[llvm-objdump] Re-commit r367284.
Add warning messages if disassembly + source for problematic inputs
Summary: Addresses https://bugs.llvm.org/show_bug.cgi?id=41905
Reviewers: jhenderson, rupprecht, grimar
Reviewed By: jhenderson, grimar
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62462
llvm-svn: 367776
Craig Topper [Sun, 4 Aug 2019 06:01:31 +0000 (06:01 +0000)]
[X86] Consistently use MVT::i8 for the constant operand of BLENDI and INSERTPS nodes.
This is the type listed in the type constraint for isel. But since
we list a type there, it doesn't get checked during isel matching.
llvm-svn: 367775
Michael Kruse [Sun, 4 Aug 2019 05:16:52 +0000 (05:16 +0000)]
Revert "[OpenMP 5.0] Codegen support for user-defined mappers."
This reverts commit r367773. The test case
OpenMP/declare_mapper_codegen.cpp is failing.
llvm-svn: 367774
Michael Kruse [Sun, 4 Aug 2019 04:18:42 +0000 (04:18 +0000)]
[OpenMP 5.0] Codegen support for user-defined mappers.
This patch implements the code generation for OpenMP 5.0 declare mapper
(user-defined mapper) constructs. For each declare mapper, a mapper
function is generated. These mapper functions will be called by the
runtime and/or other mapper functions to achieve user defined mapping.
The design slides can be found at
https://github.com/lingda-li/public-sharing/blob/master/mapper_runtime_design.pptx
Patch by Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D59474
llvm-svn: 367773
Michael Kruse [Sun, 4 Aug 2019 04:18:28 +0000 (04:18 +0000)]
[OpenMP 5.0] libomptarget interface for declare mapper functions.
This patch implements the libomptarget runtime interface for OpenMP 5.0
declare mapper functions. The declare mapper functions generated by
Clang will call them to complete the mapping of members.
kmpc_mapper_num_components gets the current number of components for a
user-defined mapper; kmpc_push_mapper_component pushes back one
component for a user-defined mapper.
The design slides can be found at
https://github.com/lingda-li/public-sharing/blob/master/mapper_runtime_design.pptx
Patch by Lingda Li <lildmh@gmail.com>
Differential Revision: https://reviews.llvm.org/D60972
llvm-svn: 367772
Craig Topper [Sun, 4 Aug 2019 02:32:06 +0000 (02:32 +0000)]
[SelectionDAG] Add node creation debug message to getMemIntrinsicNode.
llvm-svn: 367771
Eric Fiselier [Sat, 3 Aug 2019 23:54:29 +0000 (23:54 +0000)]
Suppress -Wctad-maybe-unsupported on types w/o deduction guides.
There are a handful of standard library types that are intended
to support CTAD but don't need any explicit deduction guides to
do so.
This patch adds a dummy deduction guide to those types to suppress
-Wctad-maybe-unsupported (which gets emitted in user code).
llvm-svn: 367770
Yonghong Song [Sat, 3 Aug 2019 23:41:26 +0000 (23:41 +0000)]
[Transforms] Do not drop !preserve.access.index metadata
Currently, when a GVN or CSE optimization happens,
the llvm.preserve.access.index metadata is dropped.
This caused a problem for BPF AbstructMemberOffset phase
as it relies on the metadata (debuginfo types).
This patch added proper hooks in lib/Transforms to
preserve !preserve.access.index metadata. A test
case is added to ensure metadata is preserved under CSE.
Differential Revision: https://reviews.llvm.org/D65700
llvm-svn: 367769
Craig Topper [Sat, 3 Aug 2019 23:09:13 +0000 (23:09 +0000)]
[DAGCombiner] Prevent the combine added in r367710 from creating illegal types after type legalization.
This is further fix for PR42880.
Sanjay already disabled the X86 TLI hook for non-simple types,
but we should really call isTypeLegal here if we're after type
legalization.
llvm-svn: 367768
Lang Hames [Sat, 3 Aug 2019 22:38:31 +0000 (22:38 +0000)]
[JITLink] Fix an overly-wide read in the MachO/x86-64 test case.
This should fix the build failures on some of the 32-bit bots.
llvm-svn: 367767
Sanjay Patel [Sat, 3 Aug 2019 21:46:27 +0000 (21:46 +0000)]
[x86] change free truncate hook to handle only simple types (PR42880)
This avoids the crash from:
https://bugs.llvm.org/show_bug.cgi?id=42880
...and I think it's a proper constraint for the TLI hook.
But that example raises questions about what happens to get us
into this situation (created i29 types) and what happens later
(why does legalization die on those types), so I'm not sure if
we will resolve the bug based on this change.
llvm-svn: 367766
Keno Fischer [Sat, 3 Aug 2019 21:38:19 +0000 (21:38 +0000)]
[WebAssembly] Fix allocsize attribute in sjlj lowering
Summary:
The allocsize attribute refers to call parameters by index.
Thus, when we add the extra parameter in sjlj lowering, we
need to increment the referenced paramater in the allocsize
attribute to avoid angering the Verifier.
Reviewed By: aheejin
Differential Revision: https://reviews.llvm.org/D65470
llvm-svn: 367765
Lang Hames [Sat, 3 Aug 2019 20:17:10 +0000 (20:17 +0000)]
[JITLink] Add support for MachO/x86-64 UNSIGNED relocs with length=2.
MachO/x86-64 UNSIGNED relocs are almost always 64-bit (length=3), but UNSIGNED
relocs of length=2 are allowed if the target resides in the low 32-bits. This
patch adds support for such relocations in JITLink (previously they would have
triggered an unsupported relocation error).
llvm-svn: 367764
Lang Hames [Sat, 3 Aug 2019 20:17:08 +0000 (20:17 +0000)]
[JITLink] Fix error message formatting.
llvm-svn: 367763
Eric Fiselier [Sat, 3 Aug 2019 19:03:22 +0000 (19:03 +0000)]
Revert accidental change to __member_pointer_traits_imp.
A previous patch accidentally made the primary template
an incomplete type. This broke some C++03 constructs.
llvm-svn: 367762
Mehdi Amini [Sat, 3 Aug 2019 18:53:52 +0000 (18:53 +0000)]
Fixup r367321 "Ask confirmation when `git llvm push` will push multiple commits"
I unfortunately commited an obsolete revision in r367321 that didn't
have all the changes the reviewers suggested.
llvm-svn: 367761
Hubert Tong [Sat, 3 Aug 2019 18:52:45 +0000 (18:52 +0000)]
[yaml2obj][tests] Replace 8-byte `od` conversion with 1-byte conversion
Summary:
`od` on AIX does not seem to implement 8-byte integer conversions. Work
around this by using 1-byte conversions, which can be used in this case
since the value is byte-order insensitive.
Reviewers: grimar, daltenty, xingxue, jasonliu, MaskRay
Reviewed By: grimar, MaskRay
Subscribers: MaskRay, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65671
llvm-svn: 367760
Serge Pavlov [Sat, 3 Aug 2019 16:32:49 +0000 (16:32 +0000)]
Use switch instead of series of comparisons
This is style correction, no functional changes.
Differential Revision: https://reviews.llvm.org/D65670
llvm-svn: 367759
Tim Northover [Sat, 3 Aug 2019 15:40:00 +0000 (15:40 +0000)]
IR: Cleanup after test to silence ASAN builds
llvm-svn: 367758
Stefan Stipanovic [Sat, 3 Aug 2019 15:27:41 +0000 (15:27 +0000)]
[Attributor][NFC] run clang-format on Attributor.cpp
llvm-svn: 367757
Praveen Velliengiri [Sat, 3 Aug 2019 14:42:13 +0000 (14:42 +0000)]
Speculative Compilation
[ORC] Remove Speculator Variants for Different Program Representations
[ORC] Block Freq Analysis
Speculative Compilation with Naive Block Frequency
Add Applications to OrcSpeculation
ORC v2 with Block Freq Query & Example
Deleted BenchMark Programs
Signed-off-by: preejackie <praveenvelliengiri@gmail.com>
ORCv2 comments resolved
[ORCV2] NFC
ORCv2 NFC
[ORCv2] Speculative compilation - CFGWalkQuery
ORCv2 Adapting IRSpeculationLayer to new locking scheme
llvm-svn: 367756
Tim Northover [Sat, 3 Aug 2019 14:28:34 +0000 (14:28 +0000)]
IR: print value numbers for unnamed function arguments
For consistency with normal instructions and clarity when reading IR,
it's best to print the %0, %1, ... names of function arguments in
definitions.
Also modifies the parser to accept IR in that form for obvious reasons.
llvm-svn: 367755
Sylvestre Ledru [Sat, 3 Aug 2019 13:51:58 +0000 (13:51 +0000)]
Finish moving TargetRegisterInfo::isVirtualRegister() and friends to llvm::Register as started by r367614. NFC
llvm-svn: 367754
Nikita Popov [Sat, 3 Aug 2019 06:47:23 +0000 (06:47 +0000)]
[Thumb] Fix invalid symbol redefinition due to duplicated jumptable (PR42760)
Fix for https://bugs.llvm.org/show_bug.cgi?id=42760. A tBR_JTr
instruction is duplicated by tail duplication, which results in
the same jumptable with the same label being emitted twice.
Fix this by marking tBR_JTr as not duplicable. The corresponding
ARM/Thumb instructions are already marked as not duplicable.
Additionally also mark tTBB_JT and tTBH_JT to be consistent with
Thumb2, even though this shouldn't be strictly necessary.
Differential Revision: https://reviews.llvm.org/D65606
llvm-svn: 367753
Joel E. Denny [Sat, 3 Aug 2019 06:08:19 +0000 (06:08 +0000)]
[lit] Print internal env commands
Without this patch, the internal `env` command removes `env` and its
args from the command line while parsing it. This patch modifies a
copy instead so that the original command line is printed.
Reviewed By: stella.stamenova, rnk
Differential Revision: https://reviews.llvm.org/D65624
llvm-svn: 367752
Joel E. Denny [Sat, 3 Aug 2019 06:08:04 +0000 (06:08 +0000)]
[lit] Actually run tests for internal env command
Put the main test script in the right directory, and fix a python bug
in a local script.
Reviewed By: stella.stamenova
Differential Revision: https://reviews.llvm.org/D65623
llvm-svn: 367751
Bill Wendling [Sat, 3 Aug 2019 05:52:47 +0000 (05:52 +0000)]
Emit diagnostic if an inline asm constraint requires an immediate
Summary:
An inline asm call can result in an immediate after inlining. Therefore emit a
diagnostic here if constraint requires an immediate but one isn't supplied.
Reviewers: joerg, mgorny, efriedma, rsmith
Reviewed By: joerg
Subscribers: asb, rbar, johnrusso, simoncook, apazos, sabuasal, niosHD, zzheng, edward-jones, rogfer01, MartinMosbeck, brucehoult, the_o, PkmX, jocewei, s.egerton, MaskRay, jyknight, dylanmckay, javed.absar, fedor.sergeev, jrtc27, Jim, krytarowski, eraman, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60942
llvm-svn: 367750
Eric Fiselier [Sat, 3 Aug 2019 05:01:34 +0000 (05:01 +0000)]
Simplify <type_traits> implementations.
This patch rewrites a number of old meta-function implementations
that assumed const/volatile could not be safely applied to all types.
This is no longer the case, though for some types (Ex function types),
the const qualifier can be ignored.
The largest improvement in this patch is the reduction of is_function.
Thanks to Matt Calabrese for the improved implementation.
llvm-svn: 367749
Hideto Ueno [Sat, 3 Aug 2019 04:10:50 +0000 (04:10 +0000)]
[Attributor] Fix dereferenceable callsite argument initialization
llvm-svn: 367748
Craig Topper [Sat, 3 Aug 2019 02:54:54 +0000 (02:54 +0000)]
[InstSimplify] Add test case to show bad sign bit handling for integer abs idiom in computeKnownBits.
computeKnownBits will indicate the sign bit of abs is 0 if the
the RHS operand returned by matchSelectPattern has the nsw flag set.
For abs idioms like (X >= 0) ? X : -X, the RHS returns -X. But
we can also match ((X-Y) >= 0 ? X-Y : Y-X as abs. In this case
RHS will be the Y-X operand. According to Alive, the sign bit for
this is only 0 if both the X-Y and Y-X operands have the nsw flag.
But we're only checking the Y-X operand.
llvm-svn: 367747
Hansang Bae [Sat, 3 Aug 2019 02:38:53 +0000 (02:38 +0000)]
Add OMPT support for teams construct
This change adds OMPT support for events from teams construct.
Differential Revision: https://reviews.llvm.org/D64025
llvm-svn: 367746
Fangrui Song [Sat, 3 Aug 2019 02:26:52 +0000 (02:26 +0000)]
[ELF] Move R_*_IRELATIVE from .rel[a].plt to .rel[a].dyn unless --pack-dyn-relocs=android[+relr]
An R_*_IRELATIVE represents the address of a STT_GNU_IFUNC symbol
(redirected at runtime) which is non-preemptable and is not associated
with a canonical PLT (associated with a symbol with a section index of
SHN_UNDEF but a non-zero st_value).
.rel[a].plt [DT_JMPREL, DT_JMPREL+DT_JMPRELSZ) contains relocations that
can be lazily resolved. R_*_IRELATIVE are always eagerly resolved, so
conceptually they do not belong to .rela.plt. "iplt" is mostly a misnomer.
glibc powerpc and powerpc64 do not resolve R_*_IRELATIVE if they are in .rela.plt.
// a.o - synthesized PLT call stub has an R_*_IRELATIVE
void ifunc(); int main() { ifunc(); }
// b.o
static void real() {}
asm (".type ifunc, %gnu_indirect_function");
void *ifunc() { return ℜ }
The lld-linked executable crashes. ld.bfd places R_*_IRELATIVE in
.rela.dyn and the executable works.
glibc i386, x86_64, and aarch64 have logic
(glibc/sysdeps/*/dl-machine.h:elf_machine_lazy_rel) to eagerly resolve
R_*_IRELATIVE in .rel[a].plt so the lld-linked executable works.
Move R_*_IRELATIVE from .rel[a].plt to .rel[a].dyn to fix the crashes on
glibc powerpc/powerpc64. This also helps simplifying ifunc
implementation in FreeBSD rtld-elf powerpc64.
If --pack-dyn-relocs=android[+relr] is specified, the Android packed
dynamic relocation format is used for .rela.dyn. We cannot name
in.relaIplt ".rela.dyn" because the output section will have mixed
formats. This can be improved in the future.
Reviewed By: pcc
Differential Revision: https://reviews.llvm.org/D65651
llvm-svn: 367745
Daniel Sanders [Sat, 3 Aug 2019 02:19:23 +0000 (02:19 +0000)]
MCRegister/Register: DenseMapInfo should take `const T &`
llvm-svn: 367744
Diego Astiazaran [Sat, 3 Aug 2019 02:18:06 +0000 (02:18 +0000)]
[clang-doc] Update documentation
HTML generator has been included in clang-tools-extra release notes.
clang-doc documentation file has been updated.
Differential Revision: https://reviews.llvm.org/D65622
llvm-svn: 367743
Lang Hames [Sat, 3 Aug 2019 01:33:04 +0000 (01:33 +0000)]
[ORC] Remove some old debugging output from a unit test.
llvm-svn: 367742
Eric Christopher [Sat, 3 Aug 2019 01:12:55 +0000 (01:12 +0000)]
Temporarily Revert "[PowerPC][NFC][MachinePipeliner] Add some regression testcases"
It's breaking a number of bots, e.g.:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-bootstrap-msan/builds/13893/steps/check-llvm%20msan/logs/stdio
This reverts commit r367732.
llvm-svn: 367741
Bill Wendling [Sat, 3 Aug 2019 00:26:10 +0000 (00:26 +0000)]
Add brackets to remove warnings about ambiguous 'else's.
llvm-svn: 367740
Amara Emerson [Fri, 2 Aug 2019 23:44:24 +0000 (23:44 +0000)]
Re-commit "[GlobalISel] Add legalization support for non-power-2 loads and stores""
This is an old commit that exposed a bug in the GISel importer, which caused
non-truncating stores to be selected for truncating store patterns. Now that's
been fixed in r367737 this can go back in.
llvm-svn: 367739
Craig Topper [Fri, 2 Aug 2019 23:43:53 +0000 (23:43 +0000)]
[ScalarizeMaskedMemIntrin] Bitcast the mask to the scalar domain and use scalar bit tests for the branches for expandload/compressstore.
Same as what was done for gather/scatter/load/store in r367489.
Expandload/compressstore were delayed due to lack of constant
masking handling that has since been fixed.
llvm-svn: 367738
Amara Emerson [Fri, 2 Aug 2019 23:33:13 +0000 (23:33 +0000)]
[GlobalISel] Check LLT size matches memory size for non-truncating stores.
This was causing a bug where non-truncating stores would be selected instead of truncating ones.
Differential Revision: https://reviews.llvm.org/D64845
llvm-svn: 367737
Craig Topper [Fri, 2 Aug 2019 23:18:16 +0000 (23:18 +0000)]
[X86] Use the pointer VT for the Scale node when lowering x86 gather/scatter intrinsics.
This is consistent with the target independent intrinsic handling.
Not sure this really matters since we just pull the constant out
using getZExtValue later.
llvm-svn: 367736
Yonghong Song [Fri, 2 Aug 2019 23:16:44 +0000 (23:16 +0000)]
[BPF] Handling type conversions correctly for CO-RE
With newly added debuginfo type
metadata for preserve_array_access_index() intrinsic,
this patch did the following two things:
(1). checking validity before adding a new access index
to the access chain.
(2). calculating access byte offset in IR phase
BPFAbstractMemberAccess instead of when BTF is emitted.
For (1), the metadata provided by all preserve_*_access_index()
intrinsics are used to check whether the to-be-added type
is a proper struct/union member or array element.
For (2), with all available metadata, calculating access byte
offset becomes easier in BPFAbstractMemberAccess IR phase.
This enables us to remove the unnecessary complexity in
BTFDebug.cpp.
New tests are added for
. user explicit casting to array/structure/union
. global variable (or its dereference) as the source of base
. multi demensional arrays
. array access given a base pointer
. cases where we won't generate relocation if we cannot find
type name.
Differential Revision: https://reviews.llvm.org/D65618
llvm-svn: 367735
JF Bastien [Fri, 2 Aug 2019 23:09:01 +0000 (23:09 +0000)]
Remove support for unsupported MSVC versions
Re-land r367727 with the #if fixed.
Reviewers: rnk, lebedev.ri
Subscribers: hiraditya, jkorous, dexonsmith, lebedev.ri, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65662
llvm-svn: 367734
Reid Kleckner [Fri, 2 Aug 2019 22:55:00 +0000 (22:55 +0000)]
The MinGW linker supports response files
This affects both LLD and ld.bfd.
This isn't testable with a normal driver test with -### because those
command lines are printed before response file setup. I tested manually
and confirmed it seems to do the right thing.
llvm-svn: 367733
Jinsong Ji [Fri, 2 Aug 2019 22:27:44 +0000 (22:27 +0000)]
[PowerPC][NFC][MachinePipeliner] Add some regression testcases
Exposed by refactoring in https://reviews.llvm.org/D64665.
llvm-svn: 367732
Douglas Yung [Fri, 2 Aug 2019 22:11:49 +0000 (22:11 +0000)]
Revert Fix and test inter-procedural register allocation for ARM
This reverts r367669 (git commit
f6b00c279a5587a25876752a6ecd8da0bed959dc)
This was breaking a build bot http://lab.llvm.org:8011/builders/netbsd-amd64/builds/21233
llvm-svn: 367731
Stella Stamenova [Fri, 2 Aug 2019 22:03:12 +0000 (22:03 +0000)]
[lit] Fix 42812: lit test suite can no longer be run stand-alone
Summary:
This change updates the lit.cfg file to use llvm_config when it is available, but when it is not, it directly modifies the config object. This makes it possible to run the lit tests standalone without having built llvm (as long as the correct binaries are present in the path such as FileCheck and not).
Because the lit tests don't take a hard dependency on llvm_config, some features such as system-windows have to have definitions in lit's cfg file as well. This is a potential issue as the os features sometimes change names (for example, we went from windows to system-windows, etc.). This can cause drift between lit's tests and the rest of the llvm tests.
Reviewers: probinson, mgorny
Reviewed By: mgorny
Subscribers: delcypher, llvm-commits, asmith
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65674
llvm-svn: 367730
Diego Astiazaran [Fri, 2 Aug 2019 22:02:36 +0000 (22:02 +0000)]
[clang-doc] Add flag to continue after mapping errors
The tool used to stop execution if there was an error in the mapping
phase. It will now show the error but continue with the files that were
mapped correctly if the flag is true.
Differential revision: https://reviews.llvm.org/D65627
llvm-svn: 367729
JF Bastien [Fri, 2 Aug 2019 22:02:25 +0000 (22:02 +0000)]
Revert "Remove support for unsupported MSVC versions"
Mismatched preprocessor, I'll fix in a follow-up.
llvm-svn: 367728
JF Bastien [Fri, 2 Aug 2019 21:52:35 +0000 (21:52 +0000)]
Remove support for unsupported MSVC versions
Reviewers: rnk, lebedev.ri
Subscribers: hiraditya, jkorous, dexonsmith, lebedev.ri, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65662
llvm-svn: 367727
Shafik Yaghmour [Fri, 2 Aug 2019 21:41:50 +0000 (21:41 +0000)]
Fix ClangASTContext::CreateParameterDeclaration to not call addDecl
Summary:
The change https://reviews.llvm.org/D55575 modified ClangASTContext::CreateParameterDeclaration to call decl_ctx->addDecl(decl); this caused a regression since the existing code in DWARFASTParserClang::ParseChildParameters is called with the containing DeclContext. So when end up with cases where we are parsing a parameter for a member function and the parameter is added to the CXXRecordDecl as opposed to the CXXMethodDecl. This example is given in the regression test TestBreakpointInMemberFuncWNonPrimitiveParams.py which without this fix in a modules build leads to assert on setting a breakpoint in a member function with non primitive parameters. This scenario would be common when debugging LLDB or clang.
Differential Revision: https://reviews.llvm.org/D65414
llvm-svn: 367726
Stefan Stipanovic [Fri, 2 Aug 2019 21:31:22 +0000 (21:31 +0000)]
[Attributor] Using liveness in other attributes.
Modifying other AbstractAttributes to use Liveness AA and skip dead instructions.
Reviewers: jdoerfert, uenoku
Subscribers: hiraditya, llvm-commits
Differential revision: https://reviews.llvm.org/D65243
llvm-svn: 367725
Yonghong Song [Fri, 2 Aug 2019 21:28:28 +0000 (21:28 +0000)]
[BPF] annotate DIType metadata for builtin preseve_array_access_index()
Previously, debuginfo types are annotated to
IR builtin preserve_struct_access_index() and
preserve_union_access_index(), but not
preserve_array_access_index(). The debug info
is useful to identify the root type name which
later will be used for type comparison.
For user access without explicit type conversions,
the previous scheme works as we can ignore intermediate
compiler generated type conversions (e.g., from union types to
union members) and still generate correct access index string.
The issue comes with user explicit type conversions, e.g.,
converting an array to a structure like below:
struct t { int a; char b[40]; };
struct p { int c; int d; };
struct t *var = ...;
... __builtin_preserve_access_index(&(((struct p *)&(var->b[0]))->d)) ...
Although BPF backend can derive the type of &(var->b[0]),
explicit type annotation make checking more consistent
and less error prone.
Another benefit is for multiple dimension array handling.
For example,
struct p { int c; int d; } g[8][9][10];
... __builtin_preserve_access_index(&g[2][3][4].d) ...
It would be possible to calculate the number of "struct p"'s
before accessing its member "d" if array debug info is
available as it contains each dimension range.
This patch enables to annotate IR builtin preserve_array_access_index()
with proper debuginfo type. The unit test case and language reference
is updated as well.
Signed-off-by: Yonghong Song <yhs@fb.com>
Differential Revision: https://reviews.llvm.org/D65664
llvm-svn: 367724
Amara Emerson [Fri, 2 Aug 2019 21:15:36 +0000 (21:15 +0000)]
[AArch64][GlobalISel] Eliminate redundant G_ZEXT when the source is implicitly zext-loaded.
These cases can come up when the extending loads combiner doesn't combine a
zext(load) to a zextload op, due to some other operation being in between, which
then gets simplified at a later stage.
Differential Revision: https://reviews.llvm.org/D65360
llvm-svn: 367723
Eric Fiselier [Fri, 2 Aug 2019 21:13:38 +0000 (21:13 +0000)]
Add benchmarks to test the cost of allocator
llvm-svn: 367722
Simon Pilgrim [Fri, 2 Aug 2019 21:07:07 +0000 (21:07 +0000)]
[TargetLowering] SimplifyMultipleUseDemandedBits - don't assume INSERT_VECTOR_ELT value type is simple.
Noticed by inspection - this was copied from the X86 target equivalent where we can assume its legal/simple.
llvm-svn: 367721
Daniel Sanders [Fri, 2 Aug 2019 20:23:00 +0000 (20:23 +0000)]
Use MCRegister in MCRegisterInfo's interfaces
Summary:
As part of this, define DenseMapInfo for MCRegister (and Register while I'm at it)
Depends on D65599
Reviewers: arsenm
Subscribers: MatzeB, qcolombet, jvesely, wdng, nhaehnle, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65605
llvm-svn: 367719
Philip Reames [Fri, 2 Aug 2019 20:17:37 +0000 (20:17 +0000)]
[Statepoints] Fix overalignment of loads in no-realign-stack functions
This really should have been part of 366765. For some reason, I forgot to handle the corresponding load side, and the readable test cases (using deopt vs statepoints) turned out to be overly reduced. Oops.
As seen in the test change, the problem was that we were using a load with alignment expectations rather than the unaligned variant when the stack alignment was less than that prefered type alignment.
llvm-svn: 367718
Peter Collingbourne [Fri, 2 Aug 2019 20:14:58 +0000 (20:14 +0000)]
hwasan: Remove unused field CurModuleUniqueId. NFCI.
llvm-svn: 367717
Lang Hames [Fri, 2 Aug 2019 20:09:30 +0000 (20:09 +0000)]
[ORC] Remove a dead method.
llvm-svn: 367716
Craig Topper [Fri, 2 Aug 2019 20:04:34 +0000 (20:04 +0000)]
[ScalarizeMaskedMemIntrin] Add constant mask support to expandload and compressstore scalarization
This adds support for generating all the loads or stores for a constant mask into a single basic block with no conditionals.
Differential Revision: https://reviews.llvm.org/D65613
llvm-svn: 367715
Philip Reames [Fri, 2 Aug 2019 20:01:43 +0000 (20:01 +0000)]
[Test] Demonstrate a realignment bug missed in r366765
llvm-svn: 367714
Jinsong Ji [Fri, 2 Aug 2019 19:58:00 +0000 (19:58 +0000)]
[PowerPC][NFC] Enable ADT BitVectorTest
Test on ppc64le passed.
This fix https://bugs.llvm.org/show_bug.cgi?id=42702
llvm-svn: 367713
Lang Hames [Fri, 2 Aug 2019 19:43:20 +0000 (19:43 +0000)]
[ORC] Turn on symbol-flags overrides for LLJIT on Windows by default.
libObject does not apply the Exported flag to symbols in COFF object files,
which can lead to assertions when the symbol flags initially derived from
IR added to the JIT clash with the flags seen by the JIT linker. Both
RTDyldObjectLinkingLayer and ObjectLinkingLayer have a workaround for this:
they can be told to override the flags seen by the linker with the flags
attached to the materialization responsibility object that was passed down
to the linker. This patch modifies LLJIT's setup code to enable this override
by default on platforms where COFF is the default object format.
llvm-svn: 367712
Daniel Sanders [Fri, 2 Aug 2019 19:37:17 +0000 (19:37 +0000)]
Add MCRegister and use it in MCRegisterClass::contains()
Summary:
Register can cast to MCRegister and we may want to consider asserting
!isValid() || isPhysical() when expensive checks are on.
Depends on D65554
Reviewers: arsenm
Subscribers: wdng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65599
llvm-svn: 367711
Sanjay Patel [Fri, 2 Aug 2019 19:33:46 +0000 (19:33 +0000)]
[DAGCombiner] try to convert opposing shifts to casts
This reverses a questionable IR canonicalization when a truncate
is free:
sra (add (shl X, N1C), AddC), N1C -->
sext (add (trunc X to (width - N1C)), AddC')
https://rise4fun.com/Alive/slRC
More details in PR42644:
https://bugs.llvm.org/show_bug.cgi?id=42644
I limited this to pre-legalization for code simplicity because that
should be enough to reverse the IR patterns. I don't have any
evidence (no regression test diffs) that we need to try this later.
Differential Revision: https://reviews.llvm.org/D65607
llvm-svn: 367710
Hubert Tong [Fri, 2 Aug 2019 19:26:05 +0000 (19:26 +0000)]
[Driver][test] Avoid undefined grep in darwin-ld.c
Summary:
question-mark is not a BRE special character.
POSIX.1-2017 XBD Section 9.3.2 indicates that the interpretation of `\?`
as used by rC366282 is undefined. This patch uses an ERE instead.
Reviewers: rnk, daltenty, xingxue, jasonliu
Reviewed By: rnk
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D65668
llvm-svn: 367709
Erik Pilkington [Fri, 2 Aug 2019 19:25:58 +0000 (19:25 +0000)]
Remove a dead diagnostic, NFC
This was issued in Objective-C 1 mode, but we not longer support that, so this
is just unreachable.
llvm-svn: 367708
Eric Christopher [Fri, 2 Aug 2019 19:10:37 +0000 (19:10 +0000)]
Temporarily Revert "Changing representation of cv_def_range directives in Codeview debug info assembly format for better readability"
This is breaking bots and the author asked me to revert.
This reverts commit 367704.
llvm-svn: 367707
Joseph Tremoulet [Fri, 2 Aug 2019 19:06:15 +0000 (19:06 +0000)]
Use rip-relative addressing in asm test
The absolute form is an error when targeting Darwin.
llvm-svn: 367706
Rainer Orth [Fri, 2 Aug 2019 18:55:22 +0000 (18:55 +0000)]
[sanitizer_common][tests] Fix SanitizerCommon-Unit :: ./Sanitizer-*-Test/SanitizerCommon.PthreadDestructorIterations on Solaris
SanitizerCommon.PthreadDestructorIterations currently FAILs on Solaris:
[ RUN ] SanitizerCommon.PthreadDestructorIterations
/vol/llvm/src/compiler-rt/local/lib/sanitizer_common/tests/sanitizer_posix_test.cc:58: Failure
Value of: destructor_executed
Actual: true
Expected: false
[ FAILED ] SanitizerCommon.PthreadDestructorIterations (1 ms)
It turns out that destructor is called 4 times after the first call to SpawnThread, but
5 times after the second. While PTHREAD_DESTRUCTOR_ITERATIONS is 4 in
<limits.h>, the Solaris pthread_key_create(3C) man page documents
If, after all the destructors have been called for all keys with non-
null values, there are still some keys with non-null values, the
process will be repeated. POSIX requires that this process be executed
at least PTHREAD_DESTRUCTOR_ITERATIONS times. Solaris calls the
destructors repeatedly until all values with associated destructors are
NULL. Destructors that set new values can cause an infinite loop.
The patch adjusts the test case to allow for this.
Tested on x86_64-pc-solaris2.11.
Differential Revision: https://reviews.llvm.org/D65055
llvm-svn: 367705
Nilanjana Basu [Fri, 2 Aug 2019 18:44:39 +0000 (18:44 +0000)]
Changing representation of cv_def_range directives in Codeview debug info assembly format for better readability
llvm-svn: 367704
Alina Sbirlea [Fri, 2 Aug 2019 18:37:03 +0000 (18:37 +0000)]
[NewPassManager] Resolve assertion in CGSCCPassManager when CallCounts change.
Summary:
If the CallCounts change after an iteration of the DevirtSCCRepeatedPass, this is not reflected in the local CallCounts structure triggering the assertion checking the before/after sizes.
Since it is valid for the size to change and this only uses the CallCounts for the devirtualizing heuristic, keep a <Function*, CallCount> map instead, and make the devirtualizing decision using the counts for the functions that exist both before and after the pass.
Resolves PR42726.
Reviewers: chandlerc
Subscribers: mehdi_amini, jlebar, sanjoy.google, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65621
llvm-svn: 367703
George Rimar [Fri, 2 Aug 2019 18:22:46 +0000 (18:22 +0000)]
[llvm-objcopy] - Fix the strip-dwo-groups.test.
It was reported (https://reviews.llvm.org/D65273#1612246)
that this test fails if the compilation directory contain
a "debug_" substring.
This should fix it.
llvm-svn: 367702
Shafik Yaghmour [Fri, 2 Aug 2019 18:16:04 +0000 (18:16 +0000)]
[Formatters] Temporarily disable libc++ std::function formatter due to performance issue
Summary: We have been seeing increased reports of performance issue around large project and formatting std::function variables especially in functions signatures in back traces. There are some possible fixes but exploring those fixes may take time and it is better to temporarily disable the formatter due to its impact and re-enable it once we have a fix.
Differential Revision: https://reviews.llvm.org/D65666
llvm-svn: 367701
Jessica Paquette [Fri, 2 Aug 2019 18:12:53 +0000 (18:12 +0000)]
[AArch64][GlobalISel] Support the neg_addsub_shifted_imm32 pattern
Add an equivalent ComplexRendererFns function for SelectNegArithImmed. This
allows us to select immediate adds of -1 by turning them into subtracts.
Update select-binop.mir to show that the pattern works.
Differential Revision: https://reviews.llvm.org/D65460
llvm-svn: 367700
Alina Sbirlea [Fri, 2 Aug 2019 18:06:54 +0000 (18:06 +0000)]
[SimplifyCFG] Cleanup redundant conditions [NFC].
Summary:
Since the for loop iterates over BB's predecessors, the branch conditions found must have BB as one of the successors.
For an unconditional branch the successor must be BB, added `assert`.
For a conditional branch, one of the two successors must be BB, simplify `else if` to `else` and `assert`.
Sink common instructions outside the if/else block.
Reviewers: sanjoy.google
Subscribers: jlebar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65596
llvm-svn: 367699
Simon Pilgrim [Fri, 2 Aug 2019 17:52:55 +0000 (17:52 +0000)]
[AMDGPU] Regenerated saddo.ll test file for D47927
llvm-svn: 367698
Daniel Sanders [Fri, 2 Aug 2019 17:52:17 +0000 (17:52 +0000)]
Fix ARC after r367633
llvm-svn: 367697
Peter Collingbourne [Fri, 2 Aug 2019 17:43:45 +0000 (17:43 +0000)]
CodeGen: Don't follow aliases when extracting type info.
This fixes a crash in the case where the type info object is an alias
pointing to a non-zero offset within a global or is otherwise unanalyzable
by the stripPointerCasts() function. Looking through the alias is not the
right thing to do anyway for similar reasons as D65118.
Differential Revision: https://reviews.llvm.org/D65314
llvm-svn: 367696
Sanjay Patel [Fri, 2 Aug 2019 17:39:32 +0000 (17:39 +0000)]
[InstCombine] fold cmp+select using select operand equivalence
As discussed in PR42696:
https://bugs.llvm.org/show_bug.cgi?id=42696
...but won't help that case yet.
We have an odd situation where a select operand equivalence fold was
implemented in InstSimplify when it could have been done more generally
in InstCombine if we allow dropping of {nsw,nuw,exact} from a binop operand.
Here's an example:
https://rise4fun.com/Alive/Xplr
%cmp = icmp eq i32 %x,
2147483647
%add = add nsw i32 %x, 1
%sel = select i1 %cmp, i32 -
2147483648, i32 %add
=>
%sel = add i32 %x, 1
I've left the InstSimplify code in place for now, but my guess is that we'd
prefer to remove that as a follow-up to save on code duplication and
compile-time.
Differential Revision: https://reviews.llvm.org/D65576
llvm-svn: 367695
Nathan Huckleberry [Fri, 2 Aug 2019 17:18:31 +0000 (17:18 +0000)]
[clang-tidy] Adding static analyzer check to list of clang-tidy checks
Summary:
Since clang-tidy supports use of the static analyzer there
should be documentation of how to invoke the static analyzer
checks.
Reviewers: JonasToth, aaron.ballman, NoQ, Szelethus
Reviewed By: aaron.ballman
Subscribers: nickdesaulniers, lebedev.ri, jfb, NoQ, Eugene.Zelenko, xazax.hun, baloghadamsoftware, a.sidorin, Szelethus, donat.nagy, dkrupp, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D64454
llvm-svn: 367694
James Y Knight [Fri, 2 Aug 2019 17:10:04 +0000 (17:10 +0000)]
Fix git-llvm to not delete non-empty directories.
Previously, if a directory contained only other sub-directories, one
of which was being removed, git llvm would delete the parent and all
its subdirs, even though only one should've been deleted.
This error occurred in r366590, where the commit attempted to remove
lldb/packages/Python/lldbsuite/test/tools/lldb-mi, but git-llvm
erroneously removed the entire contents of
lldb/packages/Python/lldbsuite/test/tools.
This happened because "git apply" automatically removes empty
directories locally, and the absence of a local directory was
previously taken as an indication to call 'svn rm' on that
directory. However, an empty local directory does not necessarily
indicate that the directory is truly empty.
Fix that by removing directories only when they're empty on the git
side.
Differential Revision: https://reviews.llvm.org/D65416
llvm-svn: 367693
Peter Collingbourne [Fri, 2 Aug 2019 17:02:05 +0000 (17:02 +0000)]
gn build: Merge r367667.
llvm-svn: 367692
Joseph Tremoulet [Fri, 2 Aug 2019 16:53:42 +0000 (16:53 +0000)]
Fix PC adjustment in StackFrame::GetSymbolContext
Summary:
Update StackFrame::GetSymbolContext to mirror the logic in
RegisterContextLLDB::InitializeNonZerothFrame that knows not to do the
pc decrement when the given frame is a signal trap handler frame or the
parent of one, because the pc may not follow a call in these frames.
Accomplish this by adding a behaves_like_zeroth_frame field to
lldb_private::StackFrame, set to true for the zeroth frame, for
signal handler frames, and for parents of signal handler frames.
Also add logic to propagate the signal handler flag from UnwindPlan to
the FrameType on the RegisterContextLLDB it generates, and factor out a
helper to resolve symbol and address range for an Address now that we
need to invoke it in four places.
Reviewers: jasonmolenda, clayborg, jfb
Reviewed By: jasonmolenda
Subscribers: labath, dexonsmith, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D64993
llvm-svn: 367691
Fangrui Song [Fri, 2 Aug 2019 16:31:38 +0000 (16:31 +0000)]
[Sema] Disable -Wbitwise-op-parentheses and -Wlogical-op-parentheses by default
Summary:
The -Wparentheses warnings are enabled by default in clang but they are under
-Wall in gcc (gcc/c-family/c.opt). Some of the operator precedence warnings are
oftentimes criticized as noise (clang: default; gcc: -Wall). If a warning is
very controversial, it is probably not a good idea to enable it by default.
This patch disables the rather annoying ones:
-Wbitwise-op-parentheses, e.g. i & i | i
-Wlogical-op-parentheses, e.g. i && i || i
After this change:
```
* = enabled by default
-Wall
-Wparentheses
-Wlogical-op-parentheses
-Wlogical-not-parentheses*
-Wbitwise-op-parentheses
-Wshift-op-parentheses*
-Woverloaded-shift-op-parentheses*
-Wparentheses-equality*
-Wdangling-else*
```
-Woverloaded-shift-op-parentheses is typically followed by overload
resolution failure. We can instead improve the error message, and
probably delete -Woverloaded-shift-op-parentheses in the future. Keep it
for now because it gives some diagnostics.
Reviewers: akyrtzi, jyknight, rtrieu, rsmith, aaron.ballman
Reviewed By: aaron.ballman
Subscribers: dexonsmith, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D65192
llvm-svn: 367690
Paul Robinson [Fri, 2 Aug 2019 16:07:48 +0000 (16:07 +0000)]
[doc] Give a workaround for a FileCheck regex that ends in a brace.
Addresses PR42864.
llvm-svn: 367689
Teresa Johnson [Fri, 2 Aug 2019 15:49:39 +0000 (15:49 +0000)]
Use llvm-nm instead of nm in new test to unbreak Windows bot
New test added in r367679 used nm and should use llvm-nm.
llvm-svn: 367688
Ilya Biryukov [Fri, 2 Aug 2019 15:23:04 +0000 (15:23 +0000)]
[clangd] Fix a crash when presenting values for Hover
Summary:
We should pass the expression type, not a variable type when printing
the resulting value. Variable type may be different from what the
pretty-printing function expects, e.g. have references.
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: MaskRay, jkorous, arphaman, kadircet, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D65655
llvm-svn: 367687
Lang Hames [Fri, 2 Aug 2019 15:21:37 +0000 (15:21 +0000)]
[ORC] Change the locking scheme for ThreadSafeModule.
ThreadSafeModule/ThreadSafeContext are used to manage lifetimes and locking
for LLVMContexts in ORCv2. Prior to this patch contexts were locked as soon
as an associated Module was emitted (to be compiled and linked), and were not
unlocked until the emit call returned. This could lead to deadlocks if
interdependent modules that shared contexts were compiled on different threads:
when, during emission of the first module, the dependence was discovered the
second module (which would provide the required symbol) could not be emitted as
the thread emitting the first module still held the lock.
This patch eliminates this possibility by moving to a finer-grained locking
scheme. Each client holds the module lock only while they are actively operating
on it. To make this finer grained locking simpler/safer to implement this patch
removes the explicit lock method, 'getContextLock', from ThreadSafeModule and
replaces it with a new method, 'withModuleDo', that implicitly locks the context,
calls a user-supplied function object to operate on the Module, then implicitly
unlocks the context before returning the result.
ThreadSafeModule TSM = getModule(...);
size_t NumFunctions = TSM.withModuleDo(
[](Module &M) { // <- context locked before entry to lambda.
return M.size();
});
Existing ORCv2 layers that operate on ThreadSafeModules are updated to use the
new method.
This method is used to introduce Module locking into each of the existing
layers.
llvm-svn: 367686
David Candler [Fri, 2 Aug 2019 14:44:17 +0000 (14:44 +0000)]
[NFC] Test commit, corrected some spelling in comment
Test commit, corrected some spelling in comment.
Differential Revision: https://reviews.llvm.org/D65516
llvm-svn: 367685
Haojian Wu [Fri, 2 Aug 2019 14:24:02 +0000 (14:24 +0000)]
[clangd][vscode] clang-format the the extension code.
Summary:
As we are going to grow the extension in the near future, it is time to
formalize the TS code format/style of our extension (although we'd lose the
history).
We use default options of clang-format:
- 80 max line length
- 2 space indent
Reviewers: ilya-biryukov, sammccall, jvikstrom
Subscribers: MaskRay, jkorous, arphaman, kadircet, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D65657
llvm-svn: 367684
Tim Northover [Fri, 2 Aug 2019 14:09:49 +0000 (14:09 +0000)]
GlobalISel: support swiftself attribute
llvm-svn: 367683