Commit Graph

355992 Commits

Author SHA1 Message Date
Vitaly Buka fc07c1af69 [StackSafety] Delete useless test 2020-06-02 16:08:14 -07:00
Jan Korous 39fa431c8c [Analyzer][NFC] Fix markup in WebKit checkers documentation 2020-06-02 16:04:23 -07:00
Nick Desaulniers 8eda71616f [Clang][A32/T32][Linux] -O1 implies -fomit-frame-pointer
Summary:
An upgrade of LLVM for CrOS [0] containing [1] triggered a bunch of
errors related to writing to reserved registers for a Linux kernel's
arm64 compat vdso (which is a aarch32 image).

After a discussion on LKML [2], it was determined that
-f{no-}omit-frame-pointer was not being specified. Comparing GCC and
Clang [3], it becomes apparent that GCC defaults to omitting the frame
pointer implicitly when optimizations are enabled, and Clang does not.
ie. setting -O1 (or above) implies -fomit-frame-pointer. Clang was
defaulting to -fno-omit-frame-pointer implicitly unless -fomit-frame-pointer
was set explicitly.

Why this becomes a problem is that the Linux kernel's arm64 compat vdso
contains code that uses r7. r7 is used sometimes for the frame pointer
(for example, when targeting thumb (-mthumb)). See useR7AsFramePointer()
in llvm/llvm-project/llvm/lib/Target/ARM/ARMSubtarget.h. This is mostly
for legacy/compatibility reasons, and the 2019 Q4 revision of the ARM
AAPCS looks to standardize r11 as the frame pointer for aarch32, though
this is not yet implemented in LLVM.

Users that are reliant on the implicit value if unspecified when
optimizations are enabled should explicitly choose -fomit-frame-pointer
(new behavior) or -fno-omit-frame-pointer (old behavior).

[0] https://bugs.chromium.org/p/chromium/issues/detail?id=1084372
[1] https://reviews.llvm.org/D76848
[2] https://lore.kernel.org/lkml/20200526173117.155339-1-ndesaulniers@google.com/
[3] https://godbolt.org/z/0oY39t

Reviewers: kristof.beyls, psmith, danalbert, srhines, MaskRay, ostannard, efriedma

Reviewed By: psmith, danalbert, srhines, MaskRay, efriedma

Subscribers: efriedma, olista01, MaskRay, vhscampos, cfe-commits, llvm-commits, manojgupta, llozano, glider, hctim, eugenis, pcc, peter.smith, srhines

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D80828
2020-06-02 15:54:14 -07:00
Diego Caballero 8a418e5f8e [mlir][Affine] Enable fusion of loops with vector loads/stores
This patch enables affine loop fusion for loops with affine vector loads
and stores. For that, we only had to use affine memory op interfaces in
LoopFusionUtils.cpp and Utils.cpp so that vector loads and stores are
also taken into account.

Reviewed By: andydavis1, ftynse

Differential Revision: https://reviews.llvm.org/D80971
2020-06-03 01:26:22 +03:00
Jez Ng d767de44bf [lld-macho] Fix PAGEZERO=4GB errors on Windows by ensuring enum is uint64_t
It appears that MSVC doesn't resize the enum properly to fit the
constants.
2020-06-02 15:24:31 -07:00
Eric Christopher 153a24ab0f Undo initialization of TRI in CGP as this is unconditionally initialized
later.
2020-06-02 15:08:54 -07:00
Nicolas Vasilache 7de4587805 [mlir] NFC - Add AffineMinMaxOpBase getDim/SymbolOperands 2020-06-02 18:04:45 -04:00
Richard Smith 16437992ca Undo removal of test for dr777. 2020-06-02 14:19:42 -07:00
Craig Topper 961c1b5f72 [X86] Remove DeleteNode calls from PreprocessISelDAG. Rely on the RemoveDeadNodes call at the end.
Add a MadeChange flag so we don't call RemoveDeadNodes unless
something changed.
2020-06-02 14:10:20 -07:00
Reid Kleckner 3eb16fe4e9 [LLD] Have only one SpecificAllocator per type
Previously, the SpecificAllocator was a static local in the `make<T>`
function template. Using static locals is nice because they are only
constructed and registered if they are accessed. However, if there are
multiple calls to make<> with different constructor parameters, we would
get multiple static local variable instances. This is undesirable and
leads to extra memory allocations. I noticed there were two sources of
DefinedRegular allocations while checking heap profiles.
2020-06-02 14:09:09 -07:00
Craig Topper ca4bd052f6 [X86] Cleanup inconsistencies in our zext/sext vector patterns.
-Fix one place where we had a X86vzload64 but should have had
 X86vzload32.
-Make sure all patterns that have scalar_to_vector+loadi64 also
have scalar_to_vector+f64 to match 32-bit codegen.
-Add some bitcasts that were missing from patterns.
-Make sure that if we have a scalar_to_vector+load pattern
 we also have a vzload pattern.

We probably need some better canonicalization to avoid having
so many patterns.
2020-06-02 13:50:16 -07:00
Richard Smith b5f2c4e45b PR23029 / C++ DR2233: Allow expanded parameter packs to follow
parameters with default arguments.

Directly follow the wording by relaxing the AST invariant that all
parameters after one with a default arguemnt also have default
arguments, and removing the diagnostic on missing default arguments
on a pack-expanded parameter following a parameter with a default
argument.

Testing also revealed that we need to special-case explicit
specializations of templates with a pack following a parameter with a
default argument, as such explicit specializations are otherwise
impossible to write. The standard wording doesn't address this case; a
issue has been filed.

This exposed a bug where we would briefly consider a parameter to have
no default argument while we parse a delay-parsed default argument for
that parameter, which is also fixed.

Partially incorporates a patch by Raul Tambre.
2020-06-02 13:48:59 -07:00
Kadir Cetinkaya af86a10bad
[llvm] Fix unused variable warning 2020-06-02 22:46:24 +02:00
LLVM GN Syncbot 128219faf9 [gn build] Port f99d5f8c32 2020-06-02 20:36:52 +00:00
Paula Toth d5e0dfd507 [libc] Remove integration test target from check libc.
Summary:
This is failing on the asan build because we use `-nostdlib`.
I also took this opportunity to make the target name match the naming structure we've been using.

Reviewers: sivachandra

Reviewed By: sivachandra

Subscribers: mgorny, tschuett, ecnelises, libc-commits

Tags: #libc-project

Differential Revision: https://reviews.llvm.org/D81029
2020-06-02 13:32:21 -07:00
HazemAbdelhafez 915e55c910 [mlir][spirv] Add support for matrix type
This commit adds basic matrix type support to the SPIR-V dialect
including type definition, IR assembly, parsing, printing, and
(de)serialization.

Differential Revision: https://reviews.llvm.org/D80594
2020-06-02 16:30:58 -04:00
Eric Christopher 971459c3ef Fix up clang-tidy warnings around null and pointers. 2020-06-02 13:24:20 -07:00
Amy Kwan a3ada630d8 [DAGCombiner] Combine shifts into multiply-high
This patch implements a target independent DAG combine to produce multiply-high
instructions from shifts. This DAG combine will combine shifts for any type as
long as the MULH on the narrow type is legal.

For now, it is enabled on PowerPC as PowerPC is the only target that has an
implementation of the isMulhCheaperThanMulShift TLI hook introduced in
D78271.

Moreover, this DAG combine focuses on catching the pattern:
(shift (mul (ext <narrow_type>:$a to <wide_type>), (ext <narrow_type>:$b to <wide_type>)), <narrow_width>)
to produce mulhs when we have a sign-extend, and mulhu when we have
a zero-extend.

The patch performs the following checks:
- Operation is a right shift arithmetic (sra) or logical (srl)
- Input to the shift is a multiply
- Both operands to the shift are sext/zext nodes
- The extends into the multiply are both the same
- The narrow type is half the width of the wide type
- The shift amount is the width of the narrow type
- The respective mulh operation is legal

Differential Revision: https://reviews.llvm.org/D78272
2020-06-02 15:22:48 -05:00
Fangrui Song 7694b571d9 [Driver] Add multiclass OptInFlag and OptOutFlag to simplify boolean option definition
Reviewed By: dblaikie, echristo

Differential Revision: https://reviews.llvm.org/D80883
2020-06-02 13:22:12 -07:00
Nathan James 26cb70683b [NFC][ASTMatchers] StringRef-ify and Twine-ify ASTMatchers tests. 2020-06-02 21:20:58 +01:00
Jez Ng 1e1a3f67ee [lld-macho] Ensure reads from nlist_64 structs are aligned when necessary
My test refactoring in D80217 seems to have caused yaml2obj to emit
unaligned nlist_64 structs, causing ASAN'd lld to be unhappy. I don't
think this is an issue with yaml2obj though -- llvm-mc also seems to
emit unaligned nlist_64s. This diff makes lld able to safely do aligned
reads under ASAN builds while hopefully creating no overhead for regular
builds on architectures that support unaligned reads.

Reviewed By: thakis

Differential Revision: https://reviews.llvm.org/D80414
2020-06-02 13:19:38 -07:00
Jez Ng f04d1c3b90 [lld-macho] Move all tests for erroneous inputs under invalid/
For consistency.

The no-id-dylib test was originally referencing the Inputs/ folder via a
relative path. Instead of updating that path, I decided to make the test
self-contained.

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D80217
2020-06-02 13:19:38 -07:00
Jez Ng a04c133564 [lld-macho] Set __PAGEZERO size to 4GB
That's what ld64 uses for 64-bit targets. I figured it's best to make
this change sooner rather than later since a bunch of our tests are
relying on hardcoded addresses that depend on this value.

Reviewed By: smeenai

Differential Revision: https://reviews.llvm.org/D80177
2020-06-02 13:19:38 -07:00
Jez Ng df2a5778c3 [lld-macho] Error on encountering undefined symbols
... instead of silently emitting a reference to the zero address.

Reviewed By: smeenai

Differential Revision: https://reviews.llvm.org/D80169
2020-06-02 13:19:38 -07:00
Jez Ng 6f6d91867d [lld-macho] Add some relocation validation logic
I considered making a `Target::validate()` method, but I wasn't sure how
I felt about the overhead of doing yet another switch-dispatch on the
relocation type, so I put the validation in `relocateOne` instead...
might be a bit of a micro-optimization, but `relocateOne` does assume
certain things about the relocations it gets, and this error handling
makes that explicit, so it's not a totally unreasonable code
organization.

Reviewed By: smeenai

Differential Revision: https://reviews.llvm.org/D80049
2020-06-02 13:19:38 -07:00
Douglas Yung 086be9fb20 Fix test on PS4 linux bot.
Commit 301a6da8c2 changed the test and modified a CHECK
line that is inconsisent with similar lines elsewhere in the file and was causing failures
when run in slightly different configurations. This change makes the line more consistent
and should fix the bot failure.

Failure link: http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/68593
2020-06-02 20:17:02 +00:00
Thomas Lively f99d5f8c32 [WebAssembly] Eliminate range checks on br_tables
Summary:
Jump tables for most targets cannot handle out of range indices by
themselves, so LLVM emits range checks to guard the jump
tables. WebAssembly, on the other hand, implements jump tables using
the br_table instruction, which takes a default branch target as an
operand, making the range checks redundant. This patch introduces a
new MachineFunction pass in the WebAssembly backend to find and
eliminate the redundant range checks.

Reviewers: aheejin, dschuff

Subscribers: mgorny, sbc100, jgravelle-google, hiraditya, sunfish, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D80863
2020-06-02 13:14:27 -07:00
Jan Korous d61ad66050 [Analyzer][WebKit] Check record definition is available in NoUncountedMembers checker
isRefCountable asserts that the record passed as an argument has a definition available.

Fixes: https://bugs.llvm.org/show_bug.cgi?id=46142

Differential Revision: https://reviews.llvm.org/D81017
2020-06-02 13:10:36 -07:00
kamlesh kumar e31ccee1b0 [RISCV-V] Provide muldi3 builtin assembly implementation
Provides an assembly implementation of muldi3 for RISC-V, to solve bug 43388.
Since the implementation is the same as for mulsi3, that code was moved to
`riscv/int_mul_impl.inc` and is now reused by both `mulsi3.S` and `muldi3.S`.

Differential Revision: https://reviews.llvm.org/D80465
2020-06-02 21:04:55 +01:00
Thomas Lively 237be3404b [WebAssembly] Improve macro hygiene in wasm_simd128.h
Summary:
The shuffle intrinsic macros did not parenthesize usages of their
constant parameters, which could lead to incorrect results due to
operator precedence issues. This patch fixes the problem by adding the
missing paretheses.

Reviewers: aheejin

Subscribers: dschuff, sbc100, jgravelle-google, sunfish, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D80968
2020-06-02 12:55:06 -07:00
Paula Toth b836ae24a9 [libc] Add integration tests.
Summary:
This patch aims to add integration tests to check the following:
1) Header files are generated as expected.
2) Libc functions have the correct public name.
3) Libc functions have the correct return type and parameter types.
4) Symbols are exposed in the public lib.a files.

Reviewers: sivachandra, abrachet

Reviewed By: sivachandra

Subscribers: aheejin, ecnelises, dxf, mgorny, jfb, tschuett, libc-commits

Tags: #libc-project

Differential Revision: https://reviews.llvm.org/D79192
2020-06-02 12:14:07 -07:00
dstuttar 9244be7b05 [TableGen] Avoid generating switch with just default
Summary:
Switch with just default causes an MSVC warning (warning C4065: switch statement
contains 'default' but no 'case' labels).

Change-Id: I9ddeccdef93666256b5454b164b567b73b488461

Subscribers: llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D81021
2020-06-02 19:48:07 +01:00
Alexey Bataev 2f7269b677 Fix compiler crash when an expression parsed in the tentative parsing and must be claimed in the another evaluation context.
Summary:
Clang crashes when trying to finish function body. MaybeODRUseExprs is
not empty because of const static data member parsed in outer evaluation
context, upon call for isTypeIdInParens() function. It builds
annot_primary_expr, later parsed in ParseConstantExpression() in
inner constant expression evaluation context.

Reviewers: rjmccall, rsmith

Subscribers: cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D80925
2020-06-02 14:27:33 -04:00
Diego Caballero b78b98491a Update 'git push' command in GettingStarted guide
'git push' command, without any other arguments, can do different
things depending on the local configuration of Git. This patch
updates the 'git push' command with extra arguments to be more
resilient to any local configuration.

Reviewed By: mehdi_amini

Differential Revision: https://reviews.llvm.org/D79964
2020-06-02 21:25:29 +03:00
Fangrui Song 7096e04a68 [Sema] Use isAlwaysUninit for -Wuninitialized-const-reference after D79895 2020-06-02 11:25:04 -07:00
Jonas Devlieghere 5b460fb15e [llvm-dwarfdump] Print [=<offset>] after --debug-* options in help output.
Some of the --debug-* options can take an optional offset. Although the
man page does a good job of making that clear, it's much harder to
discover from the help output.

Currently the only reference to this is the following sentence:

> Where applicable these parameters take an optional =<offset> argument
> to dump only the entry at the specified offset.

This patch changes the help output from to print [=<offset>] after the
options that take an offset.

  --debug-info[=<offset>]    - Dump the .debug_info section

rdar://problem/63150066

Differential revision: https://reviews.llvm.org/D80959
2020-06-02 11:06:11 -07:00
Matt Arsenault cdd3054255 AMDGPU: Fix a test to be more stable
The chained unconditional branches can be eliminated and it's not
relevant to the test.
2020-06-02 13:47:48 -04:00
Matt Arsenault 4b1f6cdbf9 AMDGPU: Don't run indexing mode switches with exec = 0
Add mode defs rather than special casing this like some of the other
instructions.
2020-06-02 13:47:48 -04:00
Matt Arsenault 452e0d9023 AMDGPU: Don't run mode switches with exec 0
These are scalar instructions that change vector instructions, so they
should not be executed without any active lanes.

The implementation of -amdgpu-skip-threshold also seem to be backwards
from expected, since decreasing it prevents removal.
2020-06-02 13:47:48 -04:00
Jacques Pienaar 635cde6e8c [mlir] Provide defaults to make enabling dumping simpler
Differential Revision: https://reviews.llvm.org/D80818
2020-06-02 10:47:07 -07:00
Hiroshi Yamauchi 089759b96d [PGO] Enable memcmp/bcmp size value profiling.
Summary: Following up D79751.

Reviewers: davidxl

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D80578
2020-06-02 10:27:11 -07:00
Sanjay Patel 5a82dc62d2 [InstCombine] add tests for select-of-select-shuffle; NFC 2020-06-02 13:26:21 -04:00
Sanjay Patel 5b8c79ce76 [InstCombine] regenerate complete test checks; NFC 2020-06-02 13:26:21 -04:00
Konrad Kleine eaebcbc679 [lldb] NFC remove DISALLOW_COPY_AND_ASSIGN
Summary:
This is how I applied my clang-tidy check (see
https://reviews.llvm.org/D80531) in order to remove
`DISALLOW_COPY_AND_ASSIGN` and have deleted copy ctors and deleted
assignment operators instead.

```
lang=bash
grep DISALLOW_COPY_AND_ASSIGN /opt/notnfs/kkleine/llvm/lldb -r -l | sort | uniq > files

for i in $(cat files);
do
  clang-tidy \
    --checks="-*,modernize-replace-disallow-copy-and-assign-macro" \
    --format-style=LLVM \
    --header-filter=.* \
    --fix \
    -fix-errors \
    $i;
done
```

Reviewers: espindola, labath, aprantl, teemperor

Reviewed By: labath, aprantl, teemperor

Subscribers: teemperor, aprantl, labath, emaste, sbc100, aheejin, MaskRay, arphaman, usaxena95, lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D80543
2020-06-02 13:23:53 -04:00
Zequan Wu 170b6869b5 [Clang] Add a new warning to warn when passing uninitialized variables as const reference parameters to a function
Summary:
Add a new warning -Wuninitialized-const-reference as a subgroup of -Wuninitialized to address a bug filed here: https://bugs.llvm.org/show_bug.cgi?id=45624

This warning is controlled by -Wuninitialized and can be disabled by -Wno-uninitialized-const-reference.
The warning is diagnosed when passing uninitialized variables as const reference parameters to a function.

Differential Revision: https://reviews.llvm.org/D79895
2020-06-02 10:21:02 -07:00
Min-Yih Hsu 4431d64c10 Support ExtVectorType conditional operator
Extension vectors now can be used in element-wise conditional selector.
For example:
```
R[i] = C[i]? A[i] : B[i]
```
This feature was previously only enabled in OpenCL C. Now it's also
available in C. Not that it has different behaviors than GNU vectors
(i.e. __vector_size__). Extension vectors selects on signdness of the
vector. GNU vectors on the other hand do normal bool conversions. Also,
this feature is not available in C++.

Differential Revision: https://reviews.llvm.org/D80574
2020-06-02 16:35:42 +00:00
Pavel Labath de04375ac5 [lldb] Skip tests exercising DW_OP_GNU_entry_value with dsymutil
It seems that this opcode needs explicit support in dsymutil. Disable
these tests until that is implemented.
2020-06-02 18:31:15 +02:00
Louis Dionne 6f6c8a2d96 [libc++abi] Make sure we link in CrashReporterClient.a when it's present
When building the system libc++abi for Apple, we use CrashReporterClient
to provide better crash logs when calling abort(). This is exemplified by
the fact that we test for the presence of <CrashReporterClient.h> in
abort_message.cpp.

However, we must link against CrashReporterClient.a in order to get that
functionality, otherwise we get a linking error.
2020-06-02 12:23:53 -04:00
Simon Pilgrim 0beddbf2cc TypeSymbolEmitter.h - reduce includes to forward declarations. NFC. 2020-06-02 16:30:17 +01:00
David Truby 81bf1e29aa [flang] Fix release build flags.
Summary:
This patch removes the custom CMAKE_RELEASE_CXX_FLAGS variable.
This variable being set was having the effect of removing other important
Release flags, notably `-DNDEBUG`.

This patch may need to be accompanied by fixes for the macOS issues that
the removed comment mentions; I don't have a mac to test this on though so
hopefully a reviewer can help with that.

Reviewers: Andrzej, tskeith, sscalpone

Subscribers: mgorny, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D80794
2020-06-02 16:29:12 +01:00