platform/upstream/llvm.git
4 years ago[MCA] Remove verification check on MayLoad and MayStore. NFCI
Andrea Di Biagio [Wed, 5 Feb 2020 12:51:29 +0000 (12:51 +0000)]
[MCA] Remove verification check on MayLoad and MayStore. NFCI

Field NumMicroOpcodes is currently used by mca to model the number of uOPs
dispatched from the uOp-Queue to the out of order backend.  From a 'dispatch'
point of view, an instruction with zero opcodes is still valid; it simply
doesn't consume any dispatch group slots.

However, mca doesn't expect an instruction with zero uOPs to consume pipeline
resources because it is seen as a contradiction.  In practice, it only makes
sense if such an instruction is eliminated and never really executed. It may be
that mca is being too conservative here. However I believe that mca is right,
and we should probably check that inconsistency in CodeGenSchedule.cpp (when we
also verify scheduling classes in general).

This patch removes the check for MayLoad and MayStore in mca.  That check is
probably too conservative: we are already checking if a zero-uops instruction
consumes any processor resources. Note also that instructions with unmodelled
side-effects also tend to set the MayLoad/MayStore flags even if - theoretically
speaking - they might not even consume any hw resources in practice.

In future we may want to implement different checks (possibly outside of mca)
and potentially revisit the logic in mca that verifies instructions.
For that reason I have raised PR44797.

4 years agovisitINSERT_VECTOR_ELT - pull out repeated dyn_cast. NFCI.
Simon Pilgrim [Wed, 5 Feb 2020 12:32:25 +0000 (12:32 +0000)]
visitINSERT_VECTOR_ELT - pull out repeated dyn_cast. NFCI.

This always gets called at least once.

4 years ago[ARM][LowOverheadLoops] Fix loop count chain
Sam Parker [Wed, 5 Feb 2020 13:20:50 +0000 (13:20 +0000)]
[ARM][LowOverheadLoops] Fix loop count chain

Checking that the use-def chain that performs the loop count
isSafeToRemove is not sufficient because it means that we can
remove register copies that we need to restore lr to its correct
value. This change now prevents the transform from kicking in for the
'remove-elem-moves' test which needs to addressed later on.

Differential Revision: https://reviews.llvm.org/D74037

4 years ago[ARM][LowOverheadLoops] Ensure memory predication
Sam Parker [Wed, 5 Feb 2020 12:47:03 +0000 (12:47 +0000)]
[ARM][LowOverheadLoops] Ensure memory predication

While validating each MVE instruction, check that all instructions
that touch memory are somehow predicated upon the VCTP.

Differential Revision: https://reviews.llvm.org/D73616

4 years ago[lldb] Revert some change in data-formatter-objc/main.m that broke TestDataFormatterO...
Raphael Isemann [Wed, 5 Feb 2020 13:07:19 +0000 (14:07 +0100)]
[lldb] Revert some change in data-formatter-objc/main.m that broke TestDataFormatterObjCNSDate.py

Changing the date2 to an timezone independent value broke the test as the data formatters
uses the current time zone for the summary (so changing it to a time zone independent value
would again break the test in some time zones). We anyway just care about this for date2
which will be printed in a timezone-independent summary.

4 years ago[clangd] Add the missing elaborated types in FindTarget.
Haojian Wu [Wed, 5 Feb 2020 08:56:13 +0000 (09:56 +0100)]
[clangd] Add the missing elaborated types in FindTarget.

Reviewers: sammccall

Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D74025

4 years ago[lldb][NFC] Remove some unnecessary curly braces
Raphael Isemann [Wed, 5 Feb 2020 12:28:42 +0000 (13:28 +0100)]
[lldb][NFC] Remove some unnecessary curly braces

4 years ago[lldb] Introduce i386 support in NetBSD Process plugin
Michał Górny [Sun, 12 Jan 2020 21:49:36 +0000 (22:49 +0100)]
[lldb] Introduce i386 support in NetBSD Process plugin

Introduce support for i386 platform that is shared with amd64
in the same plugin.  The concept is partially based on the Linux
implementation.

The plugin tries to reuse as much code as possible.  As a result, i386
register enums are mapped into amd64 values and those are used in actual
code.  The code for accessing FPU and debug registers is shared,
although general-purpose register layouts do not match between the two
kernel APIs and need to be #ifdef-ed.

This layout will also make it possible to add support for debugging
32-bit programs on amd64 with minimal added code.

In order for this to work, I had to add missing data for debug registers
on i386.

Differential Revision: https://reviews.llvm.org/D73802

4 years ago[lldb] Make TestDataFormatterObjCCF.py pass outside California
Raphael Isemann [Wed, 5 Feb 2020 12:23:49 +0000 (13:23 +0100)]
[lldb] Make TestDataFormatterObjCCF.py pass outside California

Summary:
This test creates its dates with `NSDate dateWithNaturalLanguageString` which is deprecated and uses the current time zone of the machine to
interpret the input string. This causes that the created NSDate has a different value depending on the locale of the machine
and we hardcoded the value for California's time zone (PST) but the data formatter gives out the GMT value as a string.

This just replaces the use with the timezone-independent dateWithTimeIntervalSince1970 (which we also use in the rest of the test)
to make this pass independently of the time zone of the machine running the test.

Reviewers: mib

Reviewed By: mib

Subscribers: lldb-commits, JDevlieghere

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D74038

4 years ago[OpenCL] Fix support for cl_khr_mipmap_image_writes
Alexey Sotkin [Mon, 27 Jan 2020 09:25:03 +0000 (12:25 +0300)]
[OpenCL] Fix support for cl_khr_mipmap_image_writes

Text of the extension is available here:
https://github.com/KhronosGroup/OpenCL-Docs/blob/master/ext/cl_khr_mipmap_image.asciidoc

Patch by Ilya Mashkov

Differential Revision: https://reviews.llvm.org/D71460

4 years ago[X86] Fix missing load latencies (PR36894)
Simon Pilgrim [Wed, 5 Feb 2020 11:36:36 +0000 (11:36 +0000)]
[X86] Fix missing load latencies (PR36894)

We weren't account for load latencies in the SSE42/AES/CLMUL schedule classes

4 years agoFix MSVC signed/unsigned warning. NFCI.
Simon Pilgrim [Wed, 5 Feb 2020 10:49:43 +0000 (10:49 +0000)]
Fix MSVC signed/unsigned warning. NFCI.

4 years ago[AVR] Add disassembly tests for supported instructions
Ayke van Laethem [Mon, 3 Feb 2020 19:01:44 +0000 (20:01 +0100)]
[AVR] Add disassembly tests for supported instructions

The disassembler of the AVR backend is incomplete: most instructions do
not correctly disassemble yet.

This patch is the first in a series to add disassembly support to the
AVR backend. It starts with adding disassembler tests for instructions
that already disassemble correctly.

Differential Revision: https://reviews.llvm.org/D73911

4 years agoPartially revert c1c9819ef91aab51b5a23fb3027adac5a2f551cc
Martin Storsjö [Wed, 5 Feb 2020 11:29:48 +0000 (13:29 +0200)]
Partially revert c1c9819ef91aab51b5a23fb3027adac5a2f551cc

Revert the part of that change that broke the
test Passes/./PluginsTests/PluginsTests.LoadPlugin.

4 years ago[CMake] Add missing component dependencies, to fix building for mingw with BUILD_SHAR...
Martin Storsjö [Sat, 1 Feb 2020 20:46:27 +0000 (22:46 +0200)]
[CMake] Add missing component dependencies, to fix building for mingw with BUILD_SHARED_LIBS

Differential Revision: https://reviews.llvm.org/D73840

4 years ago[clangd] Pull out a isProtoFile function.
Haojian Wu [Wed, 5 Feb 2020 11:03:29 +0000 (12:03 +0100)]
[clangd] Pull out a isProtoFile function.

Reviewers: kadircet

Reviewed By: kadircet

Subscribers: merge_guards_bot, mgorny, ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D73780

4 years ago[lldb] Ignore type sugar in TypeSystemClang::GetPointerType
Raphael Isemann [Wed, 5 Feb 2020 10:44:28 +0000 (11:44 +0100)]
[lldb] Ignore type sugar in TypeSystemClang::GetPointerType

Summary:
Currently having a typedef for ObjC types is breaking member access in LLDB:
```
typedef NSString Str;
NSString *s; s.length; // OK
Str *s; s.length; // Causes: member reference base type 'Str *' (aka 'NSString *') is not a structure or union
```

This works for NSString as there the type building from `NSString` -> `NSString *` will correctly
build a ObjCObjectPointerType (which is necessary to make member access with a dot possible),
but for the typedef the `Str` -> `Str *` conversion will produce an incorrect PointerType. The reason
for this is that our check in TypeSystemClang::GetPointerType is not desugaring the base type,
which causes that `Str` is not recognised as a type to a `ObjCInterface` as the check only sees the
typedef sugar that was put around it. This causes that we fall back to constructing a PointerType
instead which does not allow member access with the dot operator.

This patch just changes the check to look at the desugared type instead.

Fixes rdar://17525603

Reviewers: shafik, mib

Reviewed By: mib

Subscribers: mib, JDevlieghere, lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D73952

4 years ago[AMDGPU] Fix lowering a16 image intrinsics
Sebastian Neubauer [Thu, 23 Jan 2020 13:58:38 +0000 (14:58 +0100)]
[AMDGPU] Fix lowering a16 image intrinsics

scalar_to_vector takes only one argument, not two.
The a16 tests now also check the packing of coordinates into registers

Differential Revision: https://reviews.llvm.org/D73482

4 years ago[AMDGPU] Use v3f32 type in image instructions
Sebastian Neubauer [Wed, 29 Jan 2020 13:04:56 +0000 (14:04 +0100)]
[AMDGPU] Use v3f32 type in image instructions

This should lower the amount of used registers for gfx9.

I updated some of the changed tests with the update script because
changing them by hand is tedious.

Differential Revision: https://reviews.llvm.org/D73884

4 years ago[yaml2obj][obj2yaml] - Simplify format of the SHT_LLVM_ADDRSIG section.
Georgii Rymar [Mon, 3 Feb 2020 14:10:36 +0000 (17:10 +0300)]
[yaml2obj][obj2yaml] - Simplify format of the SHT_LLVM_ADDRSIG section.

Previously the description allowed to describe symbols with use of
`Name` and `Index` keys. This patch removes them and now it is still
possible to use either names or symbol indexes, but the code is simpler
and the format is slightly different.

Such a change will be useful for another patches, e.g:
https://reviews.llvm.org/D73788#inline-671077

Differential revision: https://reviews.llvm.org/D73888

4 years ago[clangd] Use printf instead of `echo -e` to be compliant with dash
Kadir Cetinkaya [Wed, 5 Feb 2020 09:21:26 +0000 (10:21 +0100)]
[clangd] Use printf instead of `echo -e` to be compliant with dash

4 years ago[DebugInfo] Avoid the call site param for mem instrs with multiple defs
Djordje Todorovic [Wed, 5 Feb 2020 08:50:05 +0000 (09:50 +0100)]
[DebugInfo] Avoid the call site param for mem instrs with multiple defs

We currently only handle mem instructions with a single define.
Avoid the call site parameter debug info when we find the case with
multiple defs, rather than throwing an assert.

Differential Revision: https://reviews.llvm.org/D73954

4 years ago[clangd] Preserve -nostdinc and --sysroot when calling query driver
Kadir Cetinkaya [Wed, 5 Feb 2020 08:55:56 +0000 (09:55 +0100)]
[clangd] Preserve -nostdinc and --sysroot when calling query driver

Solves this issue: https://github.com/clangd/clangd/issues/157

This is my first contribution to an llvm project, so I hope I'm doing it right!

Patch by @topisani (Tobias Pisani)!

Reviewers: kadircet, klimek

Differential Revision: https://reviews.llvm.org/D73811

4 years ago[X86] Add a DAG combine for (i32 (sext (i8 (x86isd::setcc_carry)))) -> (i32 (x86isd...
Craig Topper [Tue, 4 Feb 2020 07:35:16 +0000 (23:35 -0800)]
[X86] Add a DAG combine for (i32 (sext (i8 (x86isd::setcc_carry)))) -> (i32 (x86isd::setcc_carry)) and remove isel patterns.

Same for any_extend though we don't have coverage for that.

The test changes are because isel didn't check one use of the
setcc_carry. So in isel we would end up with two different
sized setcc_carry instructions. And since it clobbers
the flags we would need to recreate the flags for the second
instruction.

This code handles additional uses by truncating the new wide
setcc_carry back to the original size for those uses.

4 years ago[CMake] Passthrough CMAKE_SYSTEM_NAME to default builtin and runtimes target
Petr Hosek [Fri, 31 Jan 2020 22:53:07 +0000 (14:53 -0800)]
[CMake] Passthrough CMAKE_SYSTEM_NAME to default builtin and runtimes target

When building the default builtin and runtimes target, set the
CMAKE_SYSTEM_NAME to the current one. This is not necessary on
Linux and Darwin, but it appears to be necessary on Windows,
otherwise CMake fails.

Differential Revision: https://reviews.llvm.org/D73811

4 years ago[CMake][compiler-rt] Replace Windows backslashes with CMake ones
Petr Hosek [Tue, 28 Jan 2020 03:20:02 +0000 (19:20 -0800)]
[CMake][compiler-rt] Replace Windows backslashes with CMake ones

XRay builds uses llvm-config to obtain the ldflags and libs and then
passes those to CMake. Unfortunately, this breaks on Windows because
CMake tries to interpret backslashes followed by certain characters
as flags. We need to rewrite these into forward slashes that are used
by CMake (even on Windows).

Differential Revision: https://reviews.llvm.org/D73523

4 years agoRemove leftover artifacts from use of gtest.
Paula Toth [Wed, 5 Feb 2020 05:41:30 +0000 (21:41 -0800)]
Remove leftover artifacts from use of gtest.

Reviewers: abrachet

Reviewed By: abrachet

Subscribers: mgorny, MaskRay, libc-commits

Tags: #libc-project

Differential Revision: https://reviews.llvm.org/D74008

4 years agoAMDGPU/EG,CM: Implement fsqrt using recip(rsqrt(x)) instead of x * rsqrt(x)
Jan Vesely [Wed, 5 Feb 2020 00:27:19 +0000 (19:27 -0500)]
AMDGPU/EG,CM: Implement fsqrt using recip(rsqrt(x)) instead of x * rsqrt(x)

The old version might be faster on EG (RECIP_IEEE is Trans only),
but it'd need extra corner case checks.
This gives correct corner case behaviour and saves a register.
Fixes OCL CTS sqrt test (1-thread, scalar) on Turks.

Reviewer: arsenm
Differential Revision: https://reviews.llvm.org/D74017

4 years ago[WebAssembly] Add experimental multivalue calling ABI
Thomas Lively [Sat, 18 Jan 2020 02:28:43 +0000 (18:28 -0800)]
[WebAssembly] Add experimental multivalue calling ABI

Summary:
For now, this ABI simply expands all possible aggregate arguments and
returns all possible aggregates directly. This ABI will change rapidly
as we prototype and benchmark a new ABI that takes advantage of
multivalue return and possibly other changes from the MVP ABI.

Reviewers: aheejin, dschuff

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

Tags: #clang

Differential Revision: https://reviews.llvm.org/D72972

4 years agoRevert "[WebAssembly][InstrEmitter] Foundation for multivalue call lowering"
Thomas Lively [Mon, 3 Feb 2020 22:37:10 +0000 (14:37 -0800)]
Revert "[WebAssembly][InstrEmitter] Foundation for multivalue call lowering"

Summary:
This reverts commit 3ef169e586f4d14efe690c23c878d5aa92a80eb5. The
purpose of this commit was to allow stack machines to perform
instruction selection for instructions with variadic defs. However,
MachineInstrs fundamentally cannot support variadic defs right now, so
this change does not turn out to be useful.

Depends on D73927.

Reviewers: aheejin

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

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D73928

4 years agoEnhance debugserver's err reporting on attach fails
Jason Molenda [Wed, 5 Feb 2020 03:43:33 +0000 (19:43 -0800)]
Enhance debugserver's err reporting on attach fails

Explicitly check for a request to attach to a pid that doesn't
exist, to attach to a pid that is already being debugged, unify the
SIP process check, and an attempt at checking if developer mode is
enabled on the system (which isn't working in debugserver, for some
reason; I can't get the authorization record which should be an
unprivileged operation and works in a standalone program I wrote).

I'll debug the developer mode check later, but I wanted to land it
along with everything else; right now it will claim that developer
mode is always enabled so it's harmless to include as-is.

4 years ago[mlir] Mark the MLIR tools for installation in CMake
Kern Handa [Wed, 5 Feb 2020 03:42:17 +0000 (03:42 +0000)]
[mlir] Mark the MLIR tools for installation in CMake

This binplaces `mlir-translate`, `mlir-cuda-runner`, and `mlir-cpu-runner` when building the CMake install target.

Differential Revision: https://reviews.llvm.org/D73986

4 years agoAMDGPU: Correct memory size for image intrinsics
Matt Arsenault [Tue, 28 Jan 2020 15:26:52 +0000 (10:26 -0500)]
AMDGPU: Correct memory size for image intrinsics

This was incorrectly rounding up to the next power of 2. v4f32 was
rounding up to v8f32, which was just wrong. There are also v3i16/v3f16
available in MVT, so we don't even need to round the f16 cases
anymore. Additionally, this field is really an EVT so we don't even
need to consider this.

Also switch some asserts to return invalid. We should have an IR
verifier for these intrinsic return types, but for now it's better to
not assert on IR that passes the verifier.

This should also probably be fixed to consider that dmask is really
eliminating some of the loaded components.

4 years agoDebugInfo: Hash DW_OP_convert in loclists when using Split DWARF
David Blaikie [Wed, 5 Feb 2020 03:25:47 +0000 (19:25 -0800)]
DebugInfo: Hash DW_OP_convert in loclists when using Split DWARF

Originally committed in: 1ced28cbe75ff81f35ac2c71e941041eb3afcd00
            Reverted in: f75301d16d444d8cb6810d679290df744bc79ec7

(reverted due to tests failing on non-linux/x86 targets, tests have since been
generalized and specialized... since Split DWARF isn't supported on non-elf
targets anyway and we have no way to run on "whatever elf target is available"
so they fail on MacOS without an explicit target triple)

This code was incorrectly emitting extra bytes into arbitrary parts of
the object file when it was meant to be hashing them to compute the DWO
ID.

Follow-up patch(es) will refactor this API somewhat to make such bugs
harder to introduce, hopefully.

4 years agoDebugInfo: Add a couple of missing COFF sections to make convert-loclist.ll pass...
David Blaikie [Wed, 5 Feb 2020 03:23:57 +0000 (19:23 -0800)]
DebugInfo: Add a couple of missing COFF sections to make convert-loclist.ll pass on Windows

4 years agoDebugInfo: convert-debugloc.ll generalize to run on ppc64le
David Blaikie [Wed, 5 Feb 2020 03:13:42 +0000 (19:13 -0800)]
DebugInfo: convert-debugloc.ll generalize to run on ppc64le

This target produces a location list for the location, so split the
match between lines to allow for a location list match.

4 years ago[lldb/Reproducers] Change the way we instrument void* arguments
Jonas Devlieghere [Wed, 5 Feb 2020 03:00:10 +0000 (19:00 -0800)]
[lldb/Reproducers] Change the way we instrument void* arguments

The reproducer instrumentation cannot automatically serialize and
deserialize void* arguments. Currently we deal with this by explicitly
preventing these methods from being instrumented. This has the undesired
side effect of breaking replay when that method returns a value later
used by another SB API call.

The solution is to change our approach and instrument these methods.
Instead of using the DUMMY macro, we just make (de)serialization of the
void pointer a NOOP and always return a nullptr.

4 years agoDebugInfo: Fix convert-loclist.ll Split DWARF variant to use a hardcoded triple
David Blaikie [Wed, 5 Feb 2020 02:59:23 +0000 (18:59 -0800)]
DebugInfo: Fix convert-loclist.ll Split DWARF variant to use a hardcoded triple

Since we don't support Split DWARF emission on non-ELF formats, hardcode
an elfine triple (we don't have a way to ask for "any ELF triple" it
seems, so hardcoded will have to do)

4 years agoRevert "[WebAssembly] Split and recombine multivalue calls for ISel"
Thomas Lively [Mon, 3 Feb 2020 22:27:03 +0000 (14:27 -0800)]
Revert "[WebAssembly] Split and recombine multivalue calls for ISel"

Summary:
This reverts commit 28857d14a86b1e99a9d2795636a5faf17674f5a2. This
commit worked toward a solution that did not turn out to be feasible
because MachineInstrs cannot contain an arbitrary number of defs.

Reviewers: aheejin

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

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D73927

4 years ago[BPF] disable ReduceLoadWidth during SelectionDag phase
Yonghong Song [Tue, 4 Feb 2020 19:21:54 +0000 (11:21 -0800)]
[BPF] disable ReduceLoadWidth during SelectionDag phase

The compiler may transform the following code
  ctx = ctx + reloc_offset
  ... (*(u32 *)ctx) & 0x8000 ...
to
  ctx = ctx + reloc_offset
  ... (*(u8 *)(ctx + 1)) & 0x80 ...
where reloc_offset will be replaced with a constant during
AsmPrinter phase.

The above transformed code will be rejected the kernel verifier
as it does not allow
  *(type *)((ctx + non_zero_offset1) + non_zero_offset2)
style access pattern.

It is hard at SelectionDag phase to identify whether a load
is related to context or not. Sometime, interprocedure analysis
may be needed. So let us simply prevent such optimization
from happening.

Differential Revision: https://reviews.llvm.org/D73997

4 years agoRecommit: DebugInfo: Check DW_OP_convert in loclists with Split DWARF
David Blaikie [Wed, 5 Feb 2020 02:34:34 +0000 (18:34 -0800)]
Recommit: DebugInfo: Check DW_OP_convert in loclists with Split DWARF

Originally committed in: 552a8fe12bd1822f48dda2e9e8728a179f82d356
            Reverted in: f75301d16d444d8cb6810d679290df744bc79ec7

Reverted because it was running llc directly (rather than %llc_dwarf)
which uses COFF files on Windows which LLVM doesn't support all DWARF
features in.

This functionality isn't fully working, but sets up the testing for a
follow-on patch that demonstrates and fixes the brokenness related to
DWO ID hashing this construct.

4 years ago[WebAssembly] Enable recently implemented SIMD operations
Thomas Lively [Fri, 31 Jan 2020 02:23:14 +0000 (18:23 -0800)]
[WebAssembly] Enable recently implemented SIMD operations

Summary:
Moves a batch of instructions from unimplemented-simd128 to simd128
because they have recently become available in V8.

Reviewers: aheejin

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

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D73926

4 years ago[mlir][spirv] Add dialect-specific attribute for target environment
Lei Zhang [Sun, 2 Feb 2020 12:40:18 +0000 (07:40 -0500)]
[mlir][spirv] Add dialect-specific attribute for target environment

We were using normal dictionary attribute for target environment
specification. It becomes cumbersome with more and more fields.
This commit changes the modelling to a dialect-specific attribute,
where we can have control over its storage and assembly form.

Differential Revision: https://reviews.llvm.org/D73959

4 years agoPR44786: Don't assert when profiling <=> expressions.
Richard Smith [Wed, 5 Feb 2020 02:28:36 +0000 (18:28 -0800)]
PR44786: Don't assert when profiling <=> expressions.

4 years agoDebugInfo: use a symbolic DIE reference in convert-loclist.ll
David Blaikie [Wed, 5 Feb 2020 02:23:22 +0000 (18:23 -0800)]
DebugInfo: use a symbolic DIE reference in convert-loclist.ll

4 years agoReapply: DebugInfo: Add missing test coverage for DW_OP_convert in loclists
David Blaikie [Wed, 5 Feb 2020 02:08:16 +0000 (18:08 -0800)]
Reapply: DebugInfo: Add missing test coverage for DW_OP_convert in loclists

Originally committed in: 5327b917e3bd0b3db352cb5a61eea7409f2d1972
      and follow on fix: 4f281f047457ce3f1870a93253476222314f420b

            Reverted in: 191a9a78b3f4bdf35a30d3480bd630d787a2fdf6
            and: f75301d16d444d8cb6810d679290df744bc79ec7

Reverted because it wasn't portable between the targets it was running
on. Using %llc_dwarf ensures the target triple is always elfine and thus
DWARF compatible.

4 years agoDebugInfo: Generalize test/DebugInfo/X86/convert-linked.ll to run with different...
David Blaikie [Wed, 5 Feb 2020 02:02:03 +0000 (18:02 -0800)]
DebugInfo: Generalize test/DebugInfo/X86/convert-linked.ll to run with different architectures

4 years agoDebugInfo: Generalize test/DebugInfo/X86/convert-inlined.ll
David Blaikie [Wed, 5 Feb 2020 01:50:55 +0000 (17:50 -0800)]
DebugInfo: Generalize test/DebugInfo/X86/convert-inlined.ll

This test was hardcoded to only run on x86-64-linux-gnu and was overly
constrained by CHECK-NEXTing every line for an exact match.

4 years agoFix warning on trailing `;`. NFC.
Michael Liao [Wed, 5 Feb 2020 01:47:36 +0000 (20:47 -0500)]
Fix warning on trailing `;`. NFC.

4 years ago[Remarks] Fix gcc build
Francis Visoiu Mistrih [Wed, 5 Feb 2020 01:42:47 +0000 (17:42 -0800)]
[Remarks] Fix gcc build

4 years agoFix warning on trailing `;`. NFC.
Michael Liao [Wed, 5 Feb 2020 01:42:05 +0000 (20:42 -0500)]
Fix warning on trailing `;`. NFC.

4 years agoDebugInfo: convert-debugloc.ll remove erroneous CHECK
David Blaikie [Wed, 5 Feb 2020 01:39:38 +0000 (17:39 -0800)]
DebugInfo: convert-debugloc.ll remove erroneous CHECK

4 years agoDebugInfo: Generalize convert-debugloc.ll to run for multiple target architectures
David Blaikie [Wed, 5 Feb 2020 01:33:20 +0000 (17:33 -0800)]
DebugInfo: Generalize convert-debugloc.ll to run for multiple target architectures

This test was overly constrained & hardcoded only to x86-linux-gnu -
generalize the test & remove the hardcoded target triple.

4 years ago[gn build] Port b8a847c0a3e
LLVM GN Syncbot [Wed, 5 Feb 2020 01:27:20 +0000 (01:27 +0000)]
[gn build] Port b8a847c0a3e

4 years ago[gn build] Port 7531a5039fd
LLVM GN Syncbot [Wed, 5 Feb 2020 01:27:19 +0000 (01:27 +0000)]
[gn build] Port 7531a5039fd

4 years ago[cuda][hip] Temporarily XFAIL on arm
Michael Liao [Wed, 5 Feb 2020 01:25:12 +0000 (20:25 -0500)]
[cuda][hip] Temporarily XFAIL on arm

4 years agoFix BroadcasterManager::RemoveListener to really remove the listener
Reid Kleckner [Tue, 4 Feb 2020 23:56:51 +0000 (15:56 -0800)]
Fix BroadcasterManager::RemoveListener to really remove the listener

This appears to be a real bug caught by -Wunused-value. std::find_if
doesn't modify the underlying collection, it just returns an iterator
pointing to the matching element.

Reviewed By: JDevlieghere

Differential Revision: https://reviews.llvm.org/D74010

4 years ago[Remarks] Extend the RemarkStreamer to support other emitters
Francis Visoiu Mistrih [Mon, 28 Oct 2019 21:53:31 +0000 (14:53 -0700)]
[Remarks] Extend the RemarkStreamer to support other emitters

This extends the RemarkStreamer to allow for other emitters (e.g.
frontends, SIL, etc.) to emit remarks through a common interface.

See changes in llvm/docs/Remarks.rst for motivation and design choices.

Differential Revision: https://reviews.llvm.org/D73676

4 years ago[NFCI] Update according to style.
Alina Sbirlea [Wed, 5 Feb 2020 00:29:04 +0000 (16:29 -0800)]
[NFCI] Update according to style.

clang-tidy + clang-format

4 years agoFix some more -Wrange-loop-analysis warnings in AArch64TargetParser
Reid Kleckner [Wed, 5 Feb 2020 00:57:49 +0000 (16:57 -0800)]
Fix some more -Wrange-loop-analysis warnings in AArch64TargetParser

4 years ago[lldb] Fix -Wdocumentation warnings
Jonas Devlieghere [Wed, 5 Feb 2020 00:44:30 +0000 (16:44 -0800)]
[lldb] Fix -Wdocumentation warnings

4 years agoFix after c25938d
Adrian McCarthy [Tue, 4 Feb 2020 23:02:28 +0000 (15:02 -0800)]
Fix after c25938d

My refactor caused some changes in error reporting that TestAddDsymCommand.py
was checking, so this restores some of the changes to preserve the old
behavior and to un-xfail the affected test.

Differential Revision: https://reviews.llvm.org/D74001

4 years ago[Driver][Darwin] Improve tests for -nostdinc & friends
Louis Dionne [Tue, 4 Feb 2020 23:34:28 +0000 (18:34 -0500)]
[Driver][Darwin] Improve tests for -nostdinc & friends

The tests were slightly underspecifying the current behavior of Clang.

4 years ago[lldb/Reproducers] Add underscore to varaibles in LLDB_RECORD macros
Jonas Devlieghere [Wed, 5 Feb 2020 00:21:31 +0000 (16:21 -0800)]
[lldb/Reproducers] Add underscore to varaibles in LLDB_RECORD macros

Rename all variables used in the LLDB_RECORD macros to start with an
underscore to avoid conflicts with the arguments being recorded.

4 years ago[lldb/Reproducers] Change record logging to mirror replay logging
Jonas Devlieghere [Wed, 5 Feb 2020 00:10:10 +0000 (16:10 -0800)]
[lldb/Reproducers] Change record logging to mirror replay logging

Pretty printing the return value, although a good idea in theory, turned
out to be more confusing than anything else because they require
printing the log statement after the arguments and resulting value have
been serialized.

This changes the logging to behave more like how deserialization is
traced, although a bit more verbose because some of the serializers are
are calling each other.

4 years ago[X86] Add custom lowering for lrint/llrint to either cvtss2si/cvtsd2si or fist.
Craig Topper [Tue, 4 Feb 2020 23:25:16 +0000 (15:25 -0800)]
[X86] Add custom lowering for lrint/llrint to either cvtss2si/cvtsd2si or fist.

lrint/llrint are defined as rounding using the current rounding
mode. Numbers that can't be converted raise FE_INVALID and an
implementation defined value is returned. They may also write to
errno.

I believe this means we can use cvtss2si/cvtsd2si or fist to
convert as long as -fno-math-errno is passed on the command line.
Clang will leave them as libcalls if errno is enabled so they
won't become ISD::LRINT/LLRINT in SelectionDAG.

For 64-bit results on a 32-bit target we can't use cvtss2si/cvtsd2si
but we can use fist since it can write to a 64-bit memory location.
Though maybe we could consider using vcvtps2qq/vcvtpd2qq on avx512dq
targets?

gcc also does this optimization.

I think we might be able to do this with STRICT_LRINT/LLRINT as
well, but I've left that for future work.

Differential Revision: https://reviews.llvm.org/D73859

4 years agoFix x86 32bits MLIR build (NFC)
Dimitry Andric [Tue, 4 Feb 2020 23:57:02 +0000 (23:57 +0000)]
Fix x86 32bits MLIR build (NFC)

This is fixing a build error:

error: non-constant-expression cannot be narrowed from type 'unsigned int' to 'Region::iterator::difference_type' (aka 'int') in initializer list

Fix pr44767

4 years agoFix -Wunused-result warnings in LLDB
Reid Kleckner [Tue, 4 Feb 2020 23:50:44 +0000 (15:50 -0800)]
Fix -Wunused-result warnings in LLDB

These try_lock usages intentionally ignore the result, as commented.

Similar to 4d23764dddc23e74ad165086d7f471a3e0e52bf8, which was reviewed
previously.

4 years ago[compiler-rt] fix OpenBSD and Solaris build with sigaltstack interception
Evgenii Stepanov [Tue, 4 Feb 2020 23:50:44 +0000 (15:50 -0800)]
[compiler-rt] fix OpenBSD and Solaris build with sigaltstack interception

Summary:
`sigaltstack` interception is implemented D73816. This updated OpenBSD and Solaris bits to fix the build errors.

Reviewers: eugenis

Reviewed By: eugenis

Subscribers: fedor.sergeev, krytarowski, emaste, eugenis, dberris, #sanitizers, llvm-commits

Tags: #sanitizers, #llvm

Patch by Igor Sugak.

Differential Revision: https://reviews.llvm.org/D73976

4 years ago[Support] Fix warnings in ARMTargetParser.cpp
Reid Kleckner [Tue, 4 Feb 2020 23:48:22 +0000 (15:48 -0800)]
[Support] Fix warnings in ARMTargetParser.cpp

4 years ago[lldb] Rename Type::ResolveClangType to Type::ResolveCompilerType
Alex Langford [Tue, 4 Feb 2020 23:40:43 +0000 (15:40 -0800)]
[lldb] Rename Type::ResolveClangType to Type::ResolveCompilerType

This is not specific to clang types. It operates on CompilerTypes in
general.

4 years ago[lldb] Delete commented-out code
Alex Langford [Tue, 4 Feb 2020 22:59:20 +0000 (14:59 -0800)]
[lldb] Delete commented-out code

4 years agoAvoid many std::tie/tuple instantiations in ASTImporter
Reid Kleckner [Tue, 4 Feb 2020 23:22:15 +0000 (15:22 -0800)]
Avoid many std::tie/tuple instantiations in ASTImporter

To factor the error checking, use importChecked instead of importSeq.
This avoids repeating the names of all of the imported child nodes once,
and allows errors to be checked with a single conditional as it is with
importSeq.

After:
  peak memory: 601.63MB
  real: 0m19.172s
  obj size: 8,352kb

Before:
  peak memory: 954.11MB
  real: 0m26.188s
  obj size: 10,000kb

The speed is not as impressive as I hoped, but the memory use reduction
is impressive, and seems worth it.

Reviewed By: martong, shafik

Differential Revision: https://reviews.llvm.org/D73675

4 years ago[X86] Give KSET0* and KSET1* pseudos the same scheduler resource usage as KXOR/KXNOR.
Craig Topper [Tue, 4 Feb 2020 20:37:19 +0000 (12:37 -0800)]
[X86] Give KSET0* and KSET1* pseudos the same scheduler resource usage as KXOR/KXNOR.

These aren't recognized as idioms by the CPU so they still use
execution resources. We just use the pseudo to force the input
register to k0.

4 years ago[SEH] Remove CATCHPAD SDNode and X86::EH_RESTORE MachineInstr
Reid Kleckner [Thu, 30 Jan 2020 23:15:57 +0000 (15:15 -0800)]
[SEH] Remove CATCHPAD SDNode and X86::EH_RESTORE MachineInstr

The CATCHPAD node mostly existed to be selected into the EH_RESTORE
instruction, which sets the frame back up when 32-bit Windows exceptions
return to the parent function. However, creating this MachineInstr early
increases the risk that other passes will come along and insert
instructions that use the stack before ESP and EBP are restored. That
happened in PR44697.

Instead of representing these in the instruction stream early, delay it
until PEI. Mark the blocks where this needs to happen as EHPads, but not
funclet entry blocks. Passes after PEI have to be careful not to hoist
instructions that can use stack across frame setup instructions, so this
should be relatively reliable.

Fixes PR44697

Reviewed By: hans

Differential Revision: https://reviews.llvm.org/D73752

4 years ago[Concepts] Add missing CXXThisScope to function template constraint substitution
Saar Raz [Tue, 4 Feb 2020 22:51:40 +0000 (00:51 +0200)]
[Concepts] Add missing CXXThisScope to function template constraint substitution

We did not have a CXXThisScope around constraint checking of functions and
function template specializations, causing a crash when checking a constraint
that had a 'this' (bug 44689).

Recommit after fixing test.

4 years ago[lldb/Reproducers] Add missing SBFile methods to the registry
Jonas Devlieghere [Tue, 4 Feb 2020 23:09:53 +0000 (15:09 -0800)]
[lldb/Reproducers] Add missing SBFile methods to the registry

SBFile::Read and SBFile::Write were missing from the registry.

4 years agoRevert "[Concepts] Add missing CXXThisScope to function template constraint substitution"
Saar Raz [Tue, 4 Feb 2020 22:58:02 +0000 (00:58 +0200)]
Revert "[Concepts] Add missing CXXThisScope to function template constraint substitution"

This reverts commit 0c67cfdb114b4c2f5c7ec374cf12118c7fa9d768 which has a broken test.

4 years ago[Concepts] Add missing CXXThisScope to function template constraint substitution
Saar Raz [Tue, 4 Feb 2020 22:51:40 +0000 (00:51 +0200)]
[Concepts] Add missing CXXThisScope to function template constraint substitution

We did not have a CXXThisScope around constraint checking of functions and
function template specializations, causing a crash when checking a constraint
that had a 'this' (bug 44689)

4 years ago[OpenMP] Add Flush directive to OpenMPIRBuilder
Kiran Chandramohan [Tue, 4 Feb 2020 21:43:40 +0000 (21:43 +0000)]
[OpenMP] Add Flush directive to OpenMPIRBuilder

Add support for Flush in the OMPIRBuilder. This patch also adds changes
to clang to use the OMPIRBuilder when '-fopenmp-enable-irbuilder'
commandline option is used.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D70712

4 years ago[lldb] Remove unused references to ClangExpressionDeclMap
Alex Langford [Tue, 4 Feb 2020 22:40:51 +0000 (14:40 -0800)]
[lldb] Remove unused references to ClangExpressionDeclMap

4 years ago[compiler-rt] [netbsd] Fix build after "implement sigaltstack interception"
Kamil Rytarowski [Tue, 4 Feb 2020 22:30:05 +0000 (23:30 +0100)]
[compiler-rt] [netbsd] Fix build after "implement sigaltstack interception"

Fallout after: https://reviews.llvm.org/D73816

4 years ago[lldb] Remove clang classes from lldb-forward.h
Alex Langford [Mon, 3 Feb 2020 23:55:18 +0000 (15:55 -0800)]
[lldb] Remove clang classes from lldb-forward.h

Summary:
lldb-forward.h is convenient in many ways, but having clang-based
class forward declarations in there makes it easy to proliferate uses of clang
outside of plugins. Removing them makes you much more conscious of when
you're using something from clang and marks where we're using things
from clang in non-plugins.

Differential Revision: https://reviews.llvm.org/D73935

4 years ago[cmake] Fix clang builds with BUILD_SHARED=ON and CLANG_LINK_CLANG_DYLIB=ON
Tom Stellard [Tue, 4 Feb 2020 21:06:54 +0000 (13:06 -0800)]
[cmake] Fix clang builds with BUILD_SHARED=ON and CLANG_LINK_CLANG_DYLIB=ON

Summary:
We were linking all the clang objects and shared libraries into
libclang-cpp.so, which was causing the command line options to be
registered twice.

Reviewers: beanz, mgorny

Reviewed By: beanz, mgorny

Subscribers: merge_guards_bot, mgorny, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D68520

4 years ago[lldb/Unittest] Fix the modules build by including TypeSystemClang.h
Jonas Devlieghere [Tue, 4 Feb 2020 22:05:15 +0000 (14:05 -0800)]
[lldb/Unittest] Fix the modules build by including TypeSystemClang.h

This should fix the modules build on the GreenDragon bot.

4 years agoRe-land "[lldb] [testsuite] generalize `DWARFASTParserClangTests` based on `DWARFExpr...
Jonas Devlieghere [Tue, 4 Feb 2020 21:59:29 +0000 (13:59 -0800)]
Re-land "[lldb] [testsuite] generalize `DWARFASTParserClangTests` based on `DWARFExpressionTest`'s YAML"

4 years ago[libomptarget] Implement wavefront functions for amdgcn
Jon Chesterfield [Tue, 4 Feb 2020 21:55:26 +0000 (21:55 +0000)]
[libomptarget] Implement wavefront functions for amdgcn

Summary: [libomptarget] Implement wavefront functions for amdgcn

Reviewers: jdoerfert, ABataev, grokos, arsenm

Reviewed By: arsenm

Subscribers: saiislam, wdng, arsenm, jvesely, openmp-commits

Tags: #openmp

Differential Revision: https://reviews.llvm.org/D73077

4 years ago[compiler-rt] Fix sanitizer_common build for FreeBSD
Dimitry Andric [Tue, 4 Feb 2020 21:50:05 +0000 (22:50 +0100)]
[compiler-rt] Fix sanitizer_common build for FreeBSD

This was broken by 28c91219c7e2, which added `struct_stack_t_sz` only to
sanitizer_platform_limits_posix.{cpp,h}.

4 years agoRevert: [lldb] [testsuite] generalize `DWARFASTParserClangTests` based on `DWARFExpre...
Jan Kratochvil [Tue, 4 Feb 2020 21:48:27 +0000 (22:48 +0100)]
Revert: [lldb] [testsuite] generalize `DWARFASTParserClangTests` based on `DWARFExpressionTest`'s YAML

It is causing a failure on OSX, to be investigated more.

Differential Revision: https://reviews.llvm.org/D73279

4 years ago[clang-tidy] Fix "expression is redundant [misc-redundant-expression]" warning. NFCI...
Simon Pilgrim [Tue, 4 Feb 2020 21:36:11 +0000 (21:36 +0000)]
[clang-tidy] Fix "expression is redundant [misc-redundant-expression]" warning. NFCI. (PR44768)

We only accept tok::TokenKind::comment enum values so no need to add other cases. Seems to be a cut+paste typo.

4 years agoAMDGPU: Fix isAlwaysUniform for simple asm SGPR results
Matt Arsenault [Tue, 4 Feb 2020 20:25:48 +0000 (15:25 -0500)]
AMDGPU: Fix isAlwaysUniform for simple asm SGPR results

We were handling the case where the result was a struct with an
extracted SGPR component, but not for the simple case.

4 years ago[lldb/Reproducers] Treat SB classes pass by value as const references
Jonas Devlieghere [Tue, 4 Feb 2020 21:18:12 +0000 (13:18 -0800)]
[lldb/Reproducers] Treat SB classes pass by value as const references

Unless trivially copyable, SB classes that are passed by value should be
treated as const references by the reproducer infrastructure and their
address should be serialized. This is already the case on the
serialization side, but not on the deserialization side. This fixes that
by removing the NotImplementedTag.

4 years ago[lldb/Reproducers] Instrument SBFile
Jonas Devlieghere [Tue, 4 Feb 2020 21:17:12 +0000 (13:17 -0800)]
[lldb/Reproducers] Instrument SBFile

Currently SBFile isn't really instrumented, which was causing trouble
when capturing and replaying the Python test suite. The class is
particularly tricky because one of its constructors takes a FileSP which
isn't instrumented. Until we have proper shadowing in place, we'll
simply always record a nullptr.

Differential revision: https://reviews.llvm.org/D73992

4 years agoFix "expression is redundant [misc-redundant-expression]" warning (PR44768)
Simon Pilgrim [Tue, 4 Feb 2020 21:23:50 +0000 (21:23 +0000)]
Fix "expression is redundant [misc-redundant-expression]" warning (PR44768)

Be more specific that getOperandConstraint should return -1 or a uint8_t value

4 years agoAMDGPU/GlobalISel: Select G_SEXT_INREG
Matt Arsenault [Sat, 25 Jan 2020 00:38:53 +0000 (19:38 -0500)]
AMDGPU/GlobalISel: Select G_SEXT_INREG

4 years agoAMDGPU/GlobalISel: Do a better job splitting 64-bit G_SEXT_INREG
Matt Arsenault [Sat, 25 Jan 2020 14:31:09 +0000 (09:31 -0500)]
AMDGPU/GlobalISel: Do a better job splitting 64-bit G_SEXT_INREG

We don't need to expand to full shifts for the > 32-bit case. This
just switches to a sext_inreg of the high half.

4 years agoAMDGPU/GlobalISel: Legalize G_SEXT_INREG
Matt Arsenault [Tue, 4 Feb 2020 21:06:34 +0000 (16:06 -0500)]
AMDGPU/GlobalISel: Legalize G_SEXT_INREG

Split the VALU 64-bit case in RegBankSelect.

4 years ago[AMDGPU] Fix infinite loop with fma combines
Austin Kerbow [Tue, 4 Feb 2020 01:08:26 +0000 (17:08 -0800)]
[AMDGPU] Fix infinite loop with fma combines

https://reviews.llvm.org/D72312 introduced an infinite loop which involves
DAGCombiner::visitFMA and AMDGPUTargetLowering::performFNegCombine.

fma( a, fneg(b), fneg(c) ) => fneg( fma (a, b, c) ) => fma( a, fneg(b), fneg(c) ) ...

This only breaks with types where 'isFNegFree' returns flase, e.g. v4f32.
Reproducing the issue also needs the attribute 'no-signed-zeros-fp-math',
and no source mods allowed on one of the users of the Op.

This fix makes changes to indicate that it is not free to negate a fma if it
has users with source mods.

Differential Revision: https://reviews.llvm.org/D73939

4 years ago[sanitizer] Add missing declarations for sigaltstack syscall wrappers.
Evgenii Stepanov [Tue, 4 Feb 2020 20:21:06 +0000 (12:21 -0800)]
[sanitizer] Add missing declarations for sigaltstack syscall wrappers.

Subscribers: #sanitizers, llvm-commits

Tags: #sanitizers, #llvm

Differential Revision: https://reviews.llvm.org/D73987

4 years agoAMDGPU/GlobalISel: Remove extension legality hacks
Matt Arsenault [Tue, 21 Jan 2020 21:47:27 +0000 (16:47 -0500)]
AMDGPU/GlobalISel: Remove extension legality hacks

The legalization has improved since this was added, and the tests
relying on this no longer need it.