Matt Arsenault [Sun, 18 Aug 2019 00:20:43 +0000 (00:20 +0000)]
AMDGPU: Disambiguate v3f16 format in load/store tables
Currently the searchable tables report the number of dwords. These
round to the same number for 3 and 4 component d16
instructions. Change this to report the number of elements so this
isn't ambiguous.
llvm-svn: 369202
Matt Arsenault [Sun, 18 Aug 2019 00:20:42 +0000 (00:20 +0000)]
TableGen: Revert changes from r369038
These aren't needed for a specific use yet, and I meant to not commit
these.
llvm-svn: 369201
Craig Topper [Sat, 17 Aug 2019 22:46:15 +0000 (22:46 +0000)]
[X86] Add a one use check to the combineStore code that handles v16i16->v16i8 truncate+store by extending to v16i32 and then emitting a v16i32->v16i8 truncstore.
This prevent us from emitting a separate truncate and a truncating
store instruction.
llvm-svn: 369200
Yonghong Song [Sat, 17 Aug 2019 22:12:00 +0000 (22:12 +0000)]
[BPF] Fix bpf llvm-objdump issues.
Commit https://reviews.llvm.org/D57939 ("[DWARF] Refactor
RelocVisitor and fix computation of SHT_RELA-typed relocation entries)
made a change for relocation resolution when operating
on an object file.
The change unfortunately broke BPF as given SymbolValue (S) and
Addent (A), previously relocation is resolved to
S + A
and after the change, it is resolved to
S
This patch fixed the issue by resolving relocation correctly.
It looks not all relocation resolution reaches here and I did not
trace down exactly when. But I do find if the object file includes
codes in two different ELF sections than default ".text",
the above bug will be triggered.
This patch included a trivial two function source code to
demonstrate this issue. The relocation for .debug_loc is resolved
incorrectly due to this and llvm-objdump cannot display source
annotated assembly.
Differential Revision: https://reviews.llvm.org/D66372
llvm-svn: 369199
Raphael Isemann [Sat, 17 Aug 2019 21:57:51 +0000 (21:57 +0000)]
[lldb][NFC] Remove unused MaterializeInitializer and MaterializeInternalVariable
llvm-svn: 369198
Roman Lebedev [Sat, 17 Aug 2019 21:35:33 +0000 (21:35 +0000)]
[NFC][InstCombine] Some tests for 'shift amount reassoc in bit test - trunc-of-lshr' (PR42399)
Finally, the fold i was looking forward to :)
The legality check is muddy, i doubt i've groked the full generalization,
but it handles all the cases i care about, and can come up with:
https://rise4fun.com/Alive/26j
https://bugs.llvm.org/show_bug.cgi?id=42399
llvm-svn: 369197
Tan S. B. [Sat, 17 Aug 2019 20:57:52 +0000 (20:57 +0000)]
[SemaDeclCXX] Allow inheriting constructor declaration to specify a cv-qualified type
Differential Revision: https://reviews.llvm.org/D47419
llvm-svn: 369196
Kristof Umann [Sat, 17 Aug 2019 16:49:54 +0000 (16:49 +0000)]
[analyzer] Turn an assert into an if condition
Shocker, turns out that terminator conditions that are binary operators
aren't always logical operators.
llvm-svn: 369195
George Rimar [Sat, 17 Aug 2019 16:07:18 +0000 (16:07 +0000)]
Recommit r369190 "[llvm-readobj/llvm-readelf] - Improve/cleanup the error reporting API."
Fix: Add a `consumeError` call removed by mistake to 'printStackSize',
this should fix the "Expected<T> must be checked before access or destruction." reported by following bot:
http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/9743/steps/stage%201%20check/logs/stdio
Original commit message:
Currently we have the following functions for error reporting:
LLVM_ATTRIBUTE_NORETURN void reportError(Twine Msg);
void reportError(Error Err, StringRef Input);
void reportWarning(Twine Msg);
void reportWarning(StringRef Input, Error Err);
void warn(llvm::Error Err);
void error(std::error_code EC);
Problems are: naming is inconsistent, arguments order is inconsistent,
some of the functions looks excessive.
After applying this patch we have:
void reportError(Error Err, StringRef Input);
void reportError(std::error_code EC, StringRef Input);
void reportWarning(Error Err, StringRef Input);
I'd be happy to remove reportError(std::error_code EC, StringRef Input) too, but it
is used by COFF heavily.
Test cases were updated, they show an improvement introduced.
Differential revision: https://reviews.llvm.org/D66286
llvm-svn: 369194
George Rimar [Sat, 17 Aug 2019 15:36:06 +0000 (15:36 +0000)]
Revert r369190, r369192 ([llvm-readobj/llvm-readelf] - Improve/cleanup the error reporting API.)
It caused multiple BB failtures:
http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/9743/steps/stage%201%20check/logs/stdio
http://lab.llvm.org:8011/builders/clang-cmake-x86_64-sde-avx512-linux/builds/26042/steps/ninja%20check%201/logs/FAIL%3A%20LLVM%3A%3Astack-sizes.test
llvm-svn: 369193
George Rimar [Sat, 17 Aug 2019 15:24:16 +0000 (15:24 +0000)]
[llvm-readobj] - An attemp to fix BB after r369191.
Few BB failed with the following error:
Command Output (stderr):
--
/home/buildbots/ppc64be-clang-lnt-test/clang-ppc64be-lnt/llvm/test/tools/llvm-readobj/stack-sizes.test:263:19: error: BADSECTION-OUT: expected string not found in input
# BADSECTION-OUT: 8 ?
^
<stdin>:4:1: note: scanning from here
^
It doesn't reproduce on ubuntu/windows I have. Also, seems many of the bots
are happy too.
This slightly reorders the code to make fouts().flush() call earlier,
like it was before the r369191.
llvm-svn: 369192
Kang Zhang [Sat, 17 Aug 2019 14:37:05 +0000 (14:37 +0000)]
[CodeGen] Do the Simple Early Return in block-placement pass to optimize the blocks
Summary:
Fix a bug of preducessors.
In `block-placement` pass, it will create some patterns for unconditional we can do the simple early retrun.
But the `early-ret` pass is before `block-placement`, we don't want to run it again.
This patch is to do the simple early return to optimize the blocks at the last of `block-placement`.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D63972
llvm-svn: 369191
George Rimar [Sat, 17 Aug 2019 14:36:40 +0000 (14:36 +0000)]
[llvm-readobj/llvm-readelf] - Improve/cleanup the error reporting API.
urrently we have the following functions for error reporting:
--
LLVM_ATTRIBUTE_NORETURN void reportError(Twine Msg);
void reportError(Error Err, StringRef Input);
void reportWarning(Twine Msg);
void reportWarning(StringRef Input, Error Err);
void warn(llvm::Error Err);
void error(std::error_code EC);
---
Problems are: naming is inconsistent, arguments order is inconsistent,
some of the functions looks excessive.
After applying this patch we have:
---
LLVM_ATTRIBUTE_NORETURN void reportError(Error Err, StringRef Input);
LLVM_ATTRIBUTE_NORETURN void reportError(std::error_code EC, StringRef Input);
void reportWarning(Error Err, StringRef Input);
---
I'd be happy to remove reportError(std::error_code EC, StringRef Input) too, but it
is used by COFF heavily.
Test cases were updated, they show an improvement introduced.
Differential revision: https://reviews.llvm.org/D66286
llvm-svn: 369190
George Rimar [Sat, 17 Aug 2019 14:23:30 +0000 (14:23 +0000)]
[test] - Remove precomiled openbsd-phdrs.elf-x86-64 objects.
There are 2 similar openbsd-phdrs.elf-x86-64 objects committed and
used in test/Object and test/tools/llvm-objdump test cases.
There is no reason to have them, we can use YAML instead. Patch does that.
Differential revision: https://reviews.llvm.org/D66342
llvm-svn: 369189
Troy A. Johnson [Sat, 17 Aug 2019 14:20:41 +0000 (14:20 +0000)]
[circular_raw_ostream] Delegate is_displayed to contained stream
raw_ostream has an is_displayed() member function that determines if the stream
is connected to a console for display or is connected to a file/pipe. By
default, is_displayed() returns false, and derived classes like raw_fd_ostream
override it. Because circular_raw_ostream wraps another stream, its result for
is_displayed() should be the same as that stream.
Differential Revision: https://reviews.llvm.org/D66026
llvm-svn: 369188
Fangrui Song [Sat, 17 Aug 2019 10:04:18 +0000 (10:04 +0000)]
[ELF] Replace local variable hasExportDynamic with config->exportDynamic. NFC
llvm-svn: 369187
Paul Walker [Sat, 17 Aug 2019 09:22:36 +0000 (09:22 +0000)]
Revert Revert [AArch64InstrInfo] Stop getInstSizeInBytes returning non-zero for meta instructions.
This reverts r369132 (git commit
19301d75f086caae1a495d267f5d0264b225942d)
llvm-svn: 369186
Paul Walker [Sat, 17 Aug 2019 09:22:28 +0000 (09:22 +0000)]
Revert [AArch64InstrInfo] Stop getInstSizeInBytes returning non-zero for meta instructions.
This reverts r369133 (git commit
2632c677f85cba1ac2aef5d68aaf8af0f5b3c944)
llvm-svn: 369185
Fangrui Song [Sat, 17 Aug 2019 06:28:03 +0000 (06:28 +0000)]
[ELF][PPC] Fix getRelExpr for R_PPC64_REL16_HI
Fixes https://github.com/ClangBuiltLinux/linux/issues/640
R_PPC64_REL16_HI was incorrectly computed as an R_ABS relocation.
rLLD368964 made it a linker failure. Change it to use R_PC to fix the
failures.
Add ppc64-reloc-rel.s for these R_PPC64_REL* tests.
llvm-svn: 369184
Troy A. Johnson [Sat, 17 Aug 2019 04:20:24 +0000 (04:20 +0000)]
[X86] Support -mlong-double-80
Add an option group for all of the -mlong-double-* options and make
-mlong-double-80 restore the default long double behavior for X86. The
motivations are that GNU accepts the -mlong-double-80 option and that complex
Makefiles often need a way of undoing earlier options. Prior to this commit, if
one chooses 64-bit or 128-bit long double for X86, there is no way to undo that
choice and restore the 80-bit behavior.
Differential Revision: https://reviews.llvm.org/D66055
llvm-svn: 369183
Diego Astiazaran [Sat, 17 Aug 2019 01:45:03 +0000 (01:45 +0000)]
[clang-doc] Fix casting not working in gcc 5.4.0
An implicit cast of std::string to llvm::SmallString<> was breaking GCC 5.4.0 builder.
A pair using llvm::SmallString<> now uses std::string.
Differential Revision: https://reviews.llvm.org/D66378
llvm-svn: 369182
Alina Sbirlea [Sat, 17 Aug 2019 01:02:12 +0000 (01:02 +0000)]
[MemorySSA] Loop passes should mark MSSA preserved when available.
This patch applies only to the new pass manager.
Currently, when MSSA Analysis is available, and pass to each loop pass, it will be preserved by that loop pass.
Hence, mark the analysis preserved based on that condition, vs the current `EnableMSSALoopDependency`. This leaves the global flag to affect only the entry point in the loop pass manager (in FunctionToLoopPassAdaptor).
llvm-svn: 369181
Petr Hosek [Sat, 17 Aug 2019 00:54:22 +0000 (00:54 +0000)]
[Fuchsia] Create the VMO during initialization, not during exit
We want to avoid doing expensive work during atexit since the process
might be terminated before we can publish the VMO and write out the
symbolizer markup, so move the VMO creation to the initialization
phase and only write data during the atexit phase.
Differential Revision: https://reviews.llvm.org/D66323
llvm-svn: 369180
Adrian Prantl [Sat, 17 Aug 2019 00:38:58 +0000 (00:38 +0000)]
Simplify code (NFC).
llvm-svn: 369179
Petr Hosek [Sat, 17 Aug 2019 00:07:26 +0000 (00:07 +0000)]
[llvm-readobj] Unwrap the value first to avoid the error
This addresses the issue introduced in r369169, we need to unwrap
the value first before we can check whether it's empty. This also
swaps the two branches to put the common path first which should
be NFC.
llvm-svn: 369177
Adrian Prantl [Fri, 16 Aug 2019 23:47:57 +0000 (23:47 +0000)]
Add LLDB dataformatters for llvm::StringRef and lldb_private::ConstString
These data formatters make the string value appear in Xcode's
variables view (and on the command line) without having to expand the
data structure.
Differential Revision: https://reviews.llvm.org/D66354
llvm-svn: 369175
Sanjay Patel [Fri, 16 Aug 2019 23:36:28 +0000 (23:36 +0000)]
Revert r367891 - "[InstCombine] combine mul+shl separated by zext"
This reverts commit
5dbb90bfe14ace30224239cac7c61a1422fa5144.
As noted in the post-commit thread for r367891, this can create
a multiply that is lowered to a libcall that may not exist.
We need to improve the backend decomposition for integer multiply
before trying to re-land this (if it's still worthwhile after
doing the backend work).
llvm-svn: 369174
Jian Cai [Fri, 16 Aug 2019 23:30:16 +0000 (23:30 +0000)]
Reland "[ARM] push LR before __gnu_mcount_nc"
This relands r369147 with fixes to unit tests.
https://reviews.llvm.org/D65019
llvm-svn: 369173
Amara Emerson [Fri, 16 Aug 2019 23:23:40 +0000 (23:23 +0000)]
[AArch64][GlobalISel] Fix an assertion during G_UNMERGE selection for s128 types.
llvm-svn: 369172
Lang Hames [Fri, 16 Aug 2019 23:20:54 +0000 (23:20 +0000)]
[ORC] Re-introduce self-dependence accidentally dropped from a unit test.
llvm-svn: 369171
Troy A. Johnson [Fri, 16 Aug 2019 23:18:22 +0000 (23:18 +0000)]
Revert "[X86] Support -mlong-double-80"
This reverts commit
250aafa2c4a1bc2395edfe8d4365545bbe56fffe.
Caused buildbot failures -- still investigating.
llvm-svn: 369170
Petr Hosek [Fri, 16 Aug 2019 23:15:40 +0000 (23:15 +0000)]
[llvm-readobj] Fallback to PT_NOTE if file doesn't have sections
This is useful when trying to read notes from stripped files and matches
the behavior of GNU readelf and eu-readelf.
Differential Revision: https://reviews.llvm.org/D66358
llvm-svn: 369169
Sanjay Patel [Fri, 16 Aug 2019 23:10:34 +0000 (23:10 +0000)]
[CodeGenPrepare] Fix use-after-free
If OptimizeExtractBits() encountered a shift instruction with no operands at all,
it would erase the instruction, but still return false.
This previously didn’t matter because its caller would always return after
processing the instruction, but https://reviews.llvm.org/D63233 changed the
function’s caller to fall through if it returned false, which would then cause
a use-after-free detectable by ASAN.
This change makes OptimizeExtractBits return true if it removes a shift
instruction with no users, terminating processing of the instruction.
Patch by: @brentdax (Brent Royal-Gordon)
Differential Revision: https://reviews.llvm.org/D66330
llvm-svn: 369168
Jordan Rupprecht [Fri, 16 Aug 2019 23:08:56 +0000 (23:08 +0000)]
Revert [X86] SimplifyDemandedVectorElts - attempt to recombine target shuffle using DemandedElts mask (reapplied)
This reverts r368662 (git commit
1a8d790cf5f89c1df718844f13e934e39bef6ef5)
The compile-time regression repro is in https://bugs.llvm.org/show_bug.cgi?id=43024
llvm-svn: 369167
Roman Lebedev [Fri, 16 Aug 2019 23:06:37 +0000 (23:06 +0000)]
[InstCombine][NFC] reuse-constant-from-select-in-icmp.ll - check branch_weights too
llvm-svn: 369166
Julian Lettner [Fri, 16 Aug 2019 22:58:13 +0000 (22:58 +0000)]
Revert "[TSan] Don't guard #include <xpc/xpc.h>"
This reverts commit
8191585b36eb4963002cce43fb2e79e7fb05c3d6.
llvm-svn: 369165
Julian Lettner [Fri, 16 Aug 2019 22:41:25 +0000 (22:41 +0000)]
[TSan] Add interceptors for os_unfair_lock
llvm-svn: 369164
Roman Lebedev [Fri, 16 Aug 2019 22:40:06 +0000 (22:40 +0000)]
[InstCombine][NFC] Revisit tests in reuse-constant-from-select-in-icmp.ll
llvm-svn: 369163
Eli Friedman [Fri, 16 Aug 2019 22:20:14 +0000 (22:20 +0000)]
[ARM] Preserve liveness in ARMConstantIslands.
We currently don't use liveness information after this point, but it can
be useful to catch bugs using -verify-machineinstrs, and optimizations
could potentially use this information in the future.
Differential Revision: https://reviews.llvm.org/D66319
llvm-svn: 369162
Richard Smith [Fri, 16 Aug 2019 22:08:39 +0000 (22:08 +0000)]
[doc] Fix some minor formatting issues.
llvm-svn: 369161
Johannes Doerfert [Fri, 16 Aug 2019 21:59:52 +0000 (21:59 +0000)]
[Attributor] Fix: Do not partially resolve returned calls.
By partially resolving returned calls we did not record that they were
not fully resolved which caused odd behavior down the line. We could
also end up with some, but not all, returned values of the callee in the
returned values map of the caller, another odd behavior we want to
avoid.
llvm-svn: 369160
Johannes Doerfert [Fri, 16 Aug 2019 21:55:01 +0000 (21:55 +0000)]
[Attributor] Fix: Make sure we set the changed flag
The flag was updated *before* we actually run the visitor callback so we
might miss updates.
llvm-svn: 369159
Johannes Doerfert [Fri, 16 Aug 2019 21:53:49 +0000 (21:53 +0000)]
[CaptureTracking] Allow null to be in either icmp operand
Summary:
Before we required the comparison against null to be "canonical", hence
null to be operand #1. This patch allows null to be in either operand,
similar to the handling of loaded globals that follows.
Reviewers: sanjoy, hfinkel, aykevl, sstefan1, uenoku
Subscribers: hiraditya, bollu, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66321
llvm-svn: 369158
Owen Pan [Fri, 16 Aug 2019 21:49:17 +0000 (21:49 +0000)]
[clang-format] Fix the bug that joins template closer and > or >>
Also fixes a buggy test case.
See PR42404
Differential Revision: https://reviews.llvm.org/D66332
llvm-svn: 369157
Sanjay Patel [Fri, 16 Aug 2019 21:37:49 +0000 (21:37 +0000)]
[CodeGenPrepare] fix RUN line settings
I'm not sure if this was running as expected with a broken triple.
llvm-svn: 369156
Johannes Doerfert [Fri, 16 Aug 2019 21:31:11 +0000 (21:31 +0000)]
[Attributor] Add all missing attribute definitions/symbols
As a preparation to "on-demand" abstract attribute generation we need
implementations for all attributes (as they can be queried and then
created on-demand where we now fail to find one).
Reviewers: uenoku, sstefan1
Subscribers: hiraditya, bollu, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66129
llvm-svn: 369155
Jonas Devlieghere [Fri, 16 Aug 2019 21:25:40 +0000 (21:25 +0000)]
[RWMutex] Simplify availability check
Check for the actual version number for the scenarios where the macOS
version isn't available (__MAC_10_12).
llvm-svn: 369154
Jonas Devlieghere [Fri, 16 Aug 2019 21:25:36 +0000 (21:25 +0000)]
[Utility] Reimplement RegularExpression on top of llvm::Regex
Originally I wanted to remove the RegularExpression class in Utility and
replace it with llvm::Regex. However, during that transition I noticed
that there are several places where need the regular expression string.
So instead I propose to keep the RegularExpression class and make it a
thin wrapper around llvm::Regex.
This patch also removes the workaround for empty regular expressions.
The result is that we are now (more or less) POSIX conformant.
Differential revision: https://reviews.llvm.org/D66174
llvm-svn: 369153
Troy A. Johnson [Fri, 16 Aug 2019 21:00:22 +0000 (21:00 +0000)]
[X86] Support -mlong-double-80
Add an option group for all of the -mlong-double-* options and make
-mlong-double-80 restore the default long double behavior for X86. The
motivations are that GNU accepts the -mlong-double-80 option and that complex
Makefiles often need a way of undoing earlier options. Prior to this commit, if
one chooses 64-bit or 128-bit long double for X86, there is no way to undo that
choice and restore the 80-bit behavior.
Differential Revision: https://reviews.llvm.org/D66055
llvm-svn: 369152
Craig Topper [Fri, 16 Aug 2019 20:50:23 +0000 (20:50 +0000)]
[X86] Use Register/MCRegister in more places in X86
This was a quick pass through some obvious places. I haven't tried the clang-tidy check.
I also replaced the zeroes in getX86SubSuperRegister with X86::NoRegister which is the real sentinel name.
Differential Revision: https://reviews.llvm.org/D66363
llvm-svn: 369151
Julian Lettner [Fri, 16 Aug 2019 20:43:09 +0000 (20:43 +0000)]
[TSan] Don't guard #include <xpc/xpc.h>
The xpc_connection_* APIs that we are intercepting are available
starting at macOS 10.7. This is old enough so that we don't need to
guard them.
llvm-svn: 369150
Jian Cai [Fri, 16 Aug 2019 20:40:21 +0000 (20:40 +0000)]
Revert "[ARM] push LR before __gnu_mcount_nc"
This reverts commit
f4cf3b959333f62b7a7b2d7771f7010c9d8da388.
llvm-svn: 369149
Troy A. Johnson [Fri, 16 Aug 2019 20:26:48 +0000 (20:26 +0000)]
[Test Commit] Fix typo in diagtool.rst
Test commit after obtaining commit access.
llvm-svn: 369148
Jian Cai [Fri, 16 Aug 2019 20:21:08 +0000 (20:21 +0000)]
[ARM] push LR before __gnu_mcount_nc
Push LR register before calling __gnu_mcount_nc as it expects the value of LR register to be the top value of
the stack on ARM32.
Differential Revision: https://reviews.llvm.org/D65019
llvm-svn: 369147
Alexey Bataev [Fri, 16 Aug 2019 20:15:02 +0000 (20:15 +0000)]
[OPENMP5.0]Diagnose global variables in lambda not marked as declare
target.
According to OpenMP 5.0, if a lambda declaration and definition appears between a declare target directive and the matching end declare target directive, all variables that are captured by the lambda expression must also appear in a to clause.
llvm-svn: 369146
Richard Smith [Fri, 16 Aug 2019 19:53:22 +0000 (19:53 +0000)]
Stop abusing SuppressAllDiagnostics when speculatively determining
whether an expression would be valid during error recovery.
llvm-svn: 369145
Johannes Doerfert [Fri, 16 Aug 2019 19:51:23 +0000 (19:51 +0000)]
[Attributor] Towards a more structured deduction pattern
Summary:
This is the first commit aiming to structure the attribute deduction.
The base idea is that we have default propagation patterns as listed
below on top of which we can add specific, e.g., context sensitive,
logic.
Deduction patterns used in this patch:
- argument states are determined from call site argument states,
see AAAlignArgument and AAArgumentFromCallSiteArguments.
- call site argument states are determined as if they were floating
values, see AAAlignCallSiteArgument and AAAlignFloating.
- floating value states are determined by traversing the def-use chain
and combining the states determined for the leaves, see
AAAlignFloating and genericValueTraversal.
- call site return states are determined from function return states,
see AAAlignCallSiteReturned and AACallSiteReturnedFromReturned.
- function return states are determined from returned value states,
see AAAlignReturned and AAReturnedFromReturnedValues.
Through this strategy all logic for alignment is concentrated in the
AAAlignFloating::updateImpl method.
Note: This commit works on its own but is part of a larger change that
involves "on-demand" creation of abstract attributes that will
participate in the fixpoint iteration. Without this part, we sometimes
do not have an AAAlign abstract attribute to query, loosing information
we determined before. All tests have appropriate FIXMEs and the
information will be recovered once we added all parts.
Reviewers: sstefan1, uenoku
Subscribers: hiraditya, bollu, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66126
llvm-svn: 369144
Johannes Doerfert [Fri, 16 Aug 2019 19:49:00 +0000 (19:49 +0000)]
[Attributor][NFC] Introduce aliases for call site attributes
Until we have call site specific liveness and/or value information there
is no need to do call site specific deduction. Though, we need the
symbols in follow up patches that make Attributor::getAAFor return a
reference.
llvm-svn: 369143
Johannes Doerfert [Fri, 16 Aug 2019 19:36:17 +0000 (19:36 +0000)]
[Attributor] Introduce initialize calls and move code to keep attributes concise
Summary:
This patch should not change the behavior except that the added
initialize methods might indicate an optimistic fixpoint earlier. The
code movement is done to keep the attribute definitions in a single
block where it makes sense. No functional changes intended there.
Reviewers: uenoku, sstefan1
Subscribers: hiraditya, bollu, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66258
llvm-svn: 369142
Lang Hames [Fri, 16 Aug 2019 19:33:37 +0000 (19:33 +0000)]
[ORC] Remove some stray debugging output accidentally left in r368707
llvm-svn: 369141
Sanjay Patel [Fri, 16 Aug 2019 18:51:30 +0000 (18:51 +0000)]
[InstCombine] canonicalize a scalar-select-of-vectors to vector select
This pattern may arise more frequently with an enhancement to SLP vectorization suggested in PR42755:
https://bugs.llvm.org/show_bug.cgi?id=42755
...but we should handle this pattern to make things easier for the backend either way.
For all in-tree targets that I looked at, codegen for typical vector sizes looks better when we change
to a vector select, so this is safe to do without a cost model (in other words, as a target-independent
canonicalization).
For example, if the condition of the select is a scalar, we end up with something like this on x86:
vpcmpgtd %xmm0, %xmm1, %xmm0
vpextrb $12, %xmm0, %eax
testb $1, %al
jne LBB0_2
## %bb.1:
vmovaps %xmm3, %xmm2
LBB0_2:
vmovaps %xmm2, %xmm0
Rather than the splat-condition variant:
vpcmpgtd %xmm0, %xmm1, %xmm0
vpshufd $255, %xmm0, %xmm0 ## xmm0 = xmm0[3,3,3,3]
vblendvps %xmm0, %xmm2, %xmm3, %xmm0
Differential Revision: https://reviews.llvm.org/D66095
llvm-svn: 369140
Diego Astiazaran [Fri, 16 Aug 2019 18:38:11 +0000 (18:38 +0000)]
[clang-doc] Redesign of generated HTML files
The new design includes a header (contains the project name), a main section, and a footer.
The main section is divided into three subsections. Left, middle, right. The left section contains the general index, the middle contains the info's data, and the right contains the index for the info's content.
The CSS has been updated.
A flag --project-name is added.
The Attributes attribute of the TagNode struct is now a vector of pairs because these attributes should be rendered in the insertion order.
The functions (cpp and js) that converts an Index tree structure into HTML were slightly modified; the first ul tag created is now a ol tag. The inner lists are still ul.
Differential Revision: https://reviews.llvm.org/D66353
llvm-svn: 369139
Evgeniy Stepanov [Fri, 16 Aug 2019 18:23:54 +0000 (18:23 +0000)]
Escape % in printf format string.
Fixes branch-relax-block-size.mir on the ASan builder.
llvm-svn: 369138
Guanzhong Chen [Fri, 16 Aug 2019 18:21:08 +0000 (18:21 +0000)]
[WebAssembly] Forbid use of EM_ASM with setjmp/longjmp
Summary:
We tried to support EM_ASM with setjmp/longjmp in binaryen. But with dynamic
linking thrown into the mix, the code is no longer understandable and cannot
be maintained. We also discovered more bugs in the EM_ASM handling code.
To ensure maintainability and correctness of the binaryen code, EM_ASM will
no longer be supported with setjmp/longjmp. This is probably fine since the
support was added recently and haven't be published.
Reviewers: tlively, sbc100, jgravelle-google, kripken
Reviewed By: tlively, kripken
Subscribers: dschuff, hiraditya, aheejin, sunfish, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66356
llvm-svn: 369137
Simon Pilgrim [Fri, 16 Aug 2019 18:13:22 +0000 (18:13 +0000)]
[X86] resolveTargetShuffleInputs - add DemandedElts variant. NFCI.
Nothing calls this yet, everything still goes through the non (all) DemandedElts wrapper.
llvm-svn: 369136
Amara Emerson [Fri, 16 Aug 2019 18:06:53 +0000 (18:06 +0000)]
[AArch64][GlobalISel] Lower G_SHUFFLE_VECTOR with 1 elt src and 1 elt mask.
Again, it's weird that these are allowed. Since lowering support was added in
r368709 we started crashing on compiling the neon intrinsics test in the test
suite. This fixes the lowering to fold the 1 elt src/mask case into copies.
llvm-svn: 369135
Simon Pilgrim [Fri, 16 Aug 2019 17:35:08 +0000 (17:35 +0000)]
[X86] combineExtractWithShuffle - handle extract(truncate(x), 0)
Eventually we need to generalize combineExtractWithShuffle to handle all faux shuffles and handle truncate (and X86ISD::VTRUNC etc.) there, but we're not ready yet (still creates nodes on the fly, incomplete DemandedElts support, bad use of recursive Depth limit).
llvm-svn: 369134
Paul Walker [Fri, 16 Aug 2019 17:29:53 +0000 (17:29 +0000)]
[AArch64InstrInfo] Stop getInstSizeInBytes returning non-zero for meta instructions.
Recommit with fixes for mac builders.
Summary:
AArch64InstrInfo::getInstSizeInBytes is incorrectly treating meta
instructions (e.g. CFI_INSTRUCTION) as normal instructions and
giving them a size of 4.
This results in branch relaxation calculating block sizes wrong.
Branch relaxation also considers alignment and thus a single
mistake can result in later blocks being incorrectly sized even
when they themselves do not contain meta instructions.
The net result is we might not relax a branch whose destination is
not within range.
Reviewers: nickdesaulniers, peter.smith
Reviewed By: peter.smith
Subscribers: javed.absar, kristof.beyls, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66337
> llvm-svn: 369111
llvm-svn: 369133
Paul Walker [Fri, 16 Aug 2019 17:29:42 +0000 (17:29 +0000)]
Revert [AArch64InstrInfo] Stop getInstSizeInBytes returning non-zero for meta instructions.
This reverts r369111 (git commit
3ccee5f7c4087ed119dbeba537f3df1b048a4dff)
llvm-svn: 369132
Vasileios Porpodas [Fri, 16 Aug 2019 17:21:18 +0000 (17:21 +0000)]
[SLPVectorizer] Make the scheduler aware of the TreeEntry operands.
Summary:
The scheduler's dependence graph gets the use-def dependencies by accessing the operands of the instructions in a bundle. However, buildTree_rec() may change the order of the operands in TreeEntry, and the scheduler is currently not aware of this. This is not causing any functional issues currently, because reordering is restricted to the operands of a single instruction. Once we support operand reordering across multiple TreeEntries, as shown here: http://www.llvm.org/devmtg/2019-04/slides/Poster-Porpodas-Supernode_SLP.pdf , the scheduler will need to get the correct operands from TreeEntry and not from the individual instructions.
In short, this patch:
- Connects the scheduler's bundle with the corresponding TreeEntry. It introduces new TE and Lane fields in ScheduleData.
- Moves the location where the operands of the TreeEntry are initialized. This used to take place in newTreeEntry() setting one operand at a time, but is now moved pre-order just before the recursion of buildTree_rec(). This is required because the scheduler needs to access both operands of the TreeEntry in tryScheduleBundle().
- Updates the scheduler to access the instruction operands through the TreeEntry operands instead of accessing the instruction operands directly.
Reviewers: ABataev, RKSimon, dtemirbulatov, Ayal, dorit, hfinkel
Reviewed By: ABataev
Subscribers: hiraditya, llvm-commits, lebedev.ri, rcorcs
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D62432
llvm-svn: 369131
Jonas Devlieghere [Fri, 16 Aug 2019 17:19:57 +0000 (17:19 +0000)]
[ADT] Remove llvm::make_unique utility.
All uses of llvm::make_unique should have been replaced with
std::make_unique. This patch represents the last part of the migration
and removes the utility from LLVM.
Differential revision: https://reviews.llvm.org/D66259
llvm-svn: 369130
Jordan Rose [Fri, 16 Aug 2019 17:17:45 +0000 (17:17 +0000)]
Fix llvm-config support for CMake build-mode-style builds
At some point we and/or CMake changed our build-mode-style builds from
$LLVM_OBJ_ROOT/bin/$CMAKE_CFG_INTDIR/
to
$LLVM_OBJ_ROOT/$CMAKE_CFG_INTDIR/bin/
which is way easier to use. But no one updated llvm-config.
https://reviews.llvm.org/D66326
llvm-svn: 369129
Fangrui Song [Fri, 16 Aug 2019 17:01:56 +0000 (17:01 +0000)]
[ELF][Hexagon] Replace R_HEXAGON_GOT with R_GOTPLT
R_GOTPLT is relative to .got.plt since D59594. Since R_HEXAGON_GOT
relocations always have 0 r_addend, they can use R_GOTPLT instead.
Reviewed By: sidneym
Differential Revision: https://reviews.llvm.org/D66274
llvm-svn: 369128
Sanjay Patel [Fri, 16 Aug 2019 17:01:26 +0000 (17:01 +0000)]
[SLP] add tests for PR16739; NFC
llvm-svn: 369127
Simon Pilgrim [Fri, 16 Aug 2019 16:41:38 +0000 (16:41 +0000)]
[X86] Alphabetize pass initialization definitions. NFCI.
llvm-svn: 369126
Guozhi Wei [Fri, 16 Aug 2019 16:26:12 +0000 (16:26 +0000)]
[CodeGen/Analysis] Intrinsic llvm.assume should not block tail call optimization
In function Analysis.cpp:isInTailCallPosition, instructions between call and ret are checked to see if they block tail call optimization. If an instruction is an intrinsic call, only llvm.lifetime_end is allowed and other intrinsic functions block tail call. When compiling tcmalloc, we found llvm.assume between a hot function call and ret, it blocks the optimization. But llvm.assume doesn't generate instructions, it should not block tail call.
Differential Revision: https://reviews.llvm.org/D66096
llvm-svn: 369125
Krzysztof Parzyszek [Fri, 16 Aug 2019 16:16:27 +0000 (16:16 +0000)]
[Hexagon] Generate min/max instructions for 64-bit vectors
llvm-svn: 369124
Diego Astiazaran [Fri, 16 Aug 2019 16:10:32 +0000 (16:10 +0000)]
[clang-doc] Fix records in global namespace
When a Record is declared in the global namespace, clang-doc serializes
it as a child of the global namespace, so the global namespace is now
one if its parent namespaces. This namespace was not being included in
the list of namespaces of the Info causing paths to be incorrect and the
index rendered incorrectly.
Affected tests have been fixed.
Differential revision: https://reviews.llvm.org/D66298
llvm-svn: 369123
Sander de Smalen [Fri, 16 Aug 2019 15:42:28 +0000 (15:42 +0000)]
Relanding r368987 [AArch64] Change location of frame-record within callee-save area.
Changes:
There was a condition for `!NeedsFrameRecord` missing in the assert. The
assert in question has changed to:
+ assert((!RPI.isPaired() || !NeedsFrameRecord || RPI.Reg2 != AArch64::FP ||
+ RPI.Reg1 == AArch64::LR) &&
+ "FrameRecord must be allocated together with LR");
This addresses PR43016.
llvm-svn: 369122
Sid Manning [Fri, 16 Aug 2019 15:35:02 +0000 (15:35 +0000)]
[lld][Hexagon]Support HEX_32 when building shared objects
Differential Revision: https://reviews.llvm.org/D66105
llvm-svn: 369121
Evandro Menezes [Fri, 16 Aug 2019 15:33:41 +0000 (15:33 +0000)]
[InstCombine] Simplify pow(2.0, itofp(y)) to ldexp(1.0, y)
Simplify `pow(2.0, itofp(y))` to `ldexp(1.0, y)`.
Differential revision: https://reviews.llvm.org/D65979
llvm-svn: 369120
Cyndy Ishida [Fri, 16 Aug 2019 15:30:48 +0000 (15:30 +0000)]
[TextAPI] Update reader to be supported by lib/Object
Summary:
To be able to use the TextAPI/Reader for tbd file consumption (by libObject)
it gets passed a MemoryBufferRef which isn't castable to MemoryBuffer.
Updated the tests to expect that input as well.
Reviewers: ributzka, steven_wu
Reviewed By: steven_wu
Subscribers: hiraditya, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66147
llvm-svn: 369119
David Green [Fri, 16 Aug 2019 15:13:37 +0000 (15:13 +0000)]
[ARM] MVE sext of a load is free
MVE also has some sext of loads, which will be free just as scalar
instructions are.
Differential Revision: https://reviews.llvm.org/D66008
llvm-svn: 369118
Roman Lebedev [Fri, 16 Aug 2019 15:10:41 +0000 (15:10 +0000)]
[InstCombine] Shift amount reassociation in bittest: trunc-of-shl (PR42399)
Summary:
This is continuation of D63829 / https://bugs.llvm.org/show_bug.cgi?id=42399
I thought naive pattern would solve my issue, but nope, it involved truncation,
thus more folds needed.. This isn't really the fold i'm interested in,
i need trunc-of-lshr, but i'we decided to start with `shl` because it's simpler.
In this case, no extra legality checks are needed:
https://rise4fun.com/Alive/CAb
We should be careful about not increasing instruction count,
since we need to produce `zext` because `and` is done in wider type.
Reviewers: spatel, nikic, xbolva00
Reviewed By: spatel
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66057
llvm-svn: 369117
Simon Pilgrim [Fri, 16 Aug 2019 14:47:59 +0000 (14:47 +0000)]
[X86][SSE] Add shuffled load tests from PR16739
llvm-svn: 369116
Simon Pilgrim [Fri, 16 Aug 2019 14:31:45 +0000 (14:31 +0000)]
Revert rL369112 : [X86][SSE] Add shuffled load tests from PR16739
I left typos in this from a WIP copy - reverting and I'll recommit.
llvm-svn: 369115
Luis Marques [Fri, 16 Aug 2019 14:27:50 +0000 (14:27 +0000)]
[RISCV] Convert registers from unsigned to Register
Only in public interfaces that have not yet been converted should there remain
registers with unsigned type.
Differential Revision: https://reviews.llvm.org/D66252
llvm-svn: 369114
Raphael Isemann [Fri, 16 Aug 2019 14:27:35 +0000 (14:27 +0000)]
[lldb][NFC] Allow for-ranges on StringList
llvm-svn: 369113
Simon Pilgrim [Fri, 16 Aug 2019 14:26:11 +0000 (14:26 +0000)]
[X86][SSE] Add shuffled load tests from PR16739
llvm-svn: 369112
Paul Walker [Fri, 16 Aug 2019 14:17:52 +0000 (14:17 +0000)]
[AArch64InstrInfo] Stop getInstSizeInBytes returning non-zero for meta instructions.
Summary:
AArch64InstrInfo::getInstSizeInBytes is incorrectly treating meta
instructions (e.g. CFI_INSTRUCTION) as normal instructions and
giving them a size of 4.
This results in branch relaxation calculating block sizes wrong.
Branch relaxation also considers alignment and thus a single
mistake can result in later blocks being incorrectly sized even
when they themselves do not contain meta instructions.
The net result is we might not relax a branch whose destination is
not within range.
Reviewers: nickdesaulniers, peter.smith
Reviewed By: peter.smith
Subscribers: javed.absar, kristof.beyls, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66337
llvm-svn: 369111
Simon Pilgrim [Fri, 16 Aug 2019 14:05:46 +0000 (14:05 +0000)]
[X86] Remove unused include. NFCI.
We don't use anything from TargetOptions.h directly and its included via TargetLowering.h anyhow.
llvm-svn: 369110
Fangrui Song [Fri, 16 Aug 2019 13:54:44 +0000 (13:54 +0000)]
[libcxxabi] __cxa_guard_require: test guard byte with != 0 instead of == 1
llvm-svn: 369109
David Green [Fri, 16 Aug 2019 13:42:39 +0000 (13:42 +0000)]
[ARM] Correct register for narrowing and widening MVE loads and stores.
The widening and narrowing MVE instructions like VLDRH.32 are only permitted to
use low tGPR registers. This means that if they are used for a stack slot,
where the register used is only decided during frame setup, we need to be able
to correctly pick a thumb1 register over a normal GPR.
This attempts to add the required logic into eliminateFrameIndex and
rewriteT2FrameIndex, only picking the FrameReg if it is a valid register for
the operands register class, and picking a valid scratch register for the
register class.
Differential Revision: https://reviews.llvm.org/D66285
llvm-svn: 369108
Sanjay Patel [Fri, 16 Aug 2019 13:31:23 +0000 (13:31 +0000)]
[x86] fix fdiv test; NFC
The test was just added with rL369106, but forgot to update the instruction
along with the test name.
llvm-svn: 369107
Sanjay Patel [Fri, 16 Aug 2019 13:23:52 +0000 (13:23 +0000)]
[x86] add tests for fdiv with variable operands; NFC
D66050 proposes to change the estimate sequence, but we
don't seem to have test coverage for the common case.
llvm-svn: 369106
Haojian Wu [Fri, 16 Aug 2019 13:20:51 +0000 (13:20 +0000)]
[clangd] suppress -Wparentheses warning: suggest parentheses around ‘&&’ within ‘||’
llvm-svn: 369105
Florian Hahn [Fri, 16 Aug 2019 13:19:29 +0000 (13:19 +0000)]
Revert [CodeGen] Do the Simple Early Return in block-placement pass to optimize the blocks
This reverts r368997 (git commit
2a903c0b679bae1919f9fc01f78e4bc6cff2add0)
It looks like this commit adds invalid predecessors to MBBs. The example
below fails the verifier after MachineBlockPlacement (run llc
-verify-machineinstrs):
@global.4 = external constant i8*
declare i32 @zot(...)
define i16* @snork.67() personality i8* bitcast (i32 (...)* @zot to i8*) {
bb:
invoke void undef()
to label %bb5 unwind label %bb4
bb4: ; preds = %bb
%tmp = landingpad { i8*, i32 }
catch i8* null
unreachable
bb5: ; preds = %bb
%tmp6 = load i32, i32* null, align 4
%tmp7 = icmp eq i32 %tmp6, 0
br i1 %tmp7, label %bb14, label %bb8
bb8: ; preds = %bb11, %bb5
invoke void undef()
to label %bb9 unwind label %bb11
bb9: ; preds = %bb8
%tmp10 = invoke i16* undef()
to label %bb14 unwind label %bb11
bb11: ; preds = %bb9, %bb8
%tmp12 = landingpad { i8*, i32 }
cleanup
catch i8* bitcast (i8** @global.4 to i8*)
%tmp13 = icmp ult i64 undef, undef
br i1 %tmp13, label %bb8, label %bb14
bb14: ; preds = %bb11, %bb9, %bb5
%tmp15 = phi i16* [ null, %bb5 ], [ null, %bb11 ], [ %tmp10, %bb9 ]
ret i16* %tmp15
}
llvm-svn: 369104
Bjorn Pettersson [Fri, 16 Aug 2019 13:16:48 +0000 (13:16 +0000)]
[DAGCombiner] Add simple folds for SMULFIX/UMULFIX/SMULFIXSAT
Summary:
Add the following DAGCombiner folds for mulfix being
one of SMULFIX/UMULFIX/SMULFIXSAT:
(mulfix x, undef, scale) -> 0
(mulfix x, 0, scale) -> 0
Also added canonicalization of constants to RHS.
Reviewers: RKSimon, craig.topper, spatel
Reviewed By: RKSimon
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D66052
llvm-svn: 369103
Bjorn Pettersson [Fri, 16 Aug 2019 13:16:38 +0000 (13:16 +0000)]
[X86] Add test case for future MULFIX DAG combine folds. NFC
Add some test cases displaying the lack of DAG combine
folds for SMULFIX/UMULFIX/SMULFIXSAT when either
multiplicand is undef or zero.
It seems like widening vector legalization for X86 can
introduce fixed point multiplication of undef values.
So that is one way that such operations could appear
during ISel.
Multiplication with zero is probably more unlikely, and
could potentially be handled by InstCombine. But I do
not think it would hurt to do such folds in DAGCombiner.
This patch only adds the test case. The folds will be
added in a follow up patch.
llvm-svn: 369102
David Green [Fri, 16 Aug 2019 13:06:49 +0000 (13:06 +0000)]
[ARM] Don't pretend we know how to generate MVE VLDn
We don't yet know how to generate these instructions for MVE. And in the case
of VLD3, we don't even have the instruction. For the moment don't tell the
vectoriser that we have VLD4, just to end up serialising the results.
Differential Revision: https://reviews.llvm.org/D66009
llvm-svn: 369101