This extracts the LLVM distribution target handling into a support module.
Extraction will enable us to restructure the builds to support multiple
distribution configurations (e.g. developer and user) to permit us to build the
development package and the user package at once.

llvm-svn: 363440

Added the errors in r363415 but they were not used in the
RemarkStreamer.

llvm-svn: 363439

llvm-svn: 363438

The NVPTX target is a bit unusual in that it's the only target without a
disassembler, and one of three targets without an asm parser (and the
first one of those three in the gn build). NVPTX doesn't have those
because it's not a binary format.

The CMake build checks for the existence of
{AsmParser,Disassembler}/CMakeLists.txt when setting
LLVM_ENUM_ASM_PARSERS / LLVM_ENUM_DISASSEBLERS
(http://llvm-cs.pcc.me.uk/CMakeLists.txt#744). The GN build doesn't want
to hit the disk for things like this, so instead I'm adding explicit
`targets_with_asm_parsers` and `targets_with_disassemblers` lists. Since
both are needed rarely, they are defined in their own gni files.

Differential Revision: https://reviews.llvm.org/D63210

llvm-svn: 363437

Now that the cycle between MCTargetDesc and TargetInfo is gone
(see revisions 360709 360718 360722 360724 360726 360731 360733 360735 360736),
remove the dependency from TargetInfo on MCTargetDesc:tablegen. In most
targets, this makes MCTargetDesc:tablegen have just a single use, so
inline it there.

For AArch64, ARM, and RISCV there's still a similar cycle between
MCTargetDesc and Utils, so the MCTargetDesc:tablegen indirection is
still needed there.

Differential Revision: https://reviews.llvm.org/D63200

llvm-svn: 363436

Summary: This patch avoids the emission of maps for target link variables when unified memory is present.

Reviewers: ABataev, caomhin

Reviewed By: ABataev

Subscribers: guansong, jdoerfert, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D60883

llvm-svn: 363435

This is a branch opcode that takes a jump table pointer, jump table index and an
index into the table to do an indirect branch.

We pass both the table pointer and JTI to allow targets like ARM64 to more
easily use the existing jump table compression optimization without having to
walk up the block to find a paired G_JUMP_TABLE.

Differential Revision: https://reviews.llvm.org/D63159

llvm-svn: 363434

Differential Revision: https://reviews.llvm.org/D63250

llvm-svn: 363433

This fixes PR42218.

Differential Revision: https://reviews.llvm.org/D63249

llvm-svn: 363432

Differential Revision: https://reviews.llvm.org/D63248

llvm-svn: 363431

In addition to being unused and duplicating code, this was also wrong
(it didn't properly mark the operand as being potentially not odr-used).

This reinstates r363340, reverted in r363352.

llvm-svn: 363430

nullptr_t does not access memory.

We now reuse CK_NullToPointer to represent a conversion from a glvalue
of type nullptr_t to a prvalue of nullptr_t where necessary.

This reinstates r363337, reverted in r363352.

llvm-svn: 363429

C++ DR712 and others: handle non-odr-use resulting from an lvalue-to-rvalue conversion applied to a member access or similar not-quite-trivial lvalue expression.

Summary:
When a variable is named in a context where we can't directly emit a
reference to it (because we don't know for sure that it's going to be
defined, or it's from an enclosing function and not captured, or the
reference might not "work" for some reason), we emit a copy of the
variable as a global and use that for the known-to-be-read-only access.

This reinstates r363295, reverted in r363352, with a fix for PR42276:
we now produce a proper name for a non-odr-use reference to a static
constexpr data member. The name <mangled-name>.const is used in that
case; such names are reserved to the implementation for cases such as
this and should demangle nicely.

Reviewers: rjmccall

Subscribers: jdoerfert, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D63157

llvm-svn: 363428

Reverting because it breaks a green dragon build:
    http://green.lab.llvm.org/green/job/clang-stage2-Rthinlto/18208

This reverts r363289 (git commit eb88badf)

llvm-svn: 363427

Reverting because it depends on r363289, which breaks a green dragon build:
    http://green.lab.llvm.org/green/job/clang-stage2-Rthinlto/18208

This reverts r363292 (git commit 42a3fc13)

llvm-svn: 363426

Reverting because it depends on r363289, which breaks a green dragon
build:
    http://green.lab.llvm.org/green/job/clang-stage2-Rthinlto/18208

This reverts r363293 (git commit c37be296)

llvm-svn: 363425

https://reviews.llvm.org/D63302

llvm-svn: 363424

llvm-svn: 363423

and replace with an equilivent countTrailingZeros.

GCD is much more expensive than this, with repeated division.

This depends on D60823

Differential Revision: https://reviews.llvm.org/D61151

llvm-svn: 363422

llvm-svn: 363421

When cross-compiling LLVM to android from Windows (for LLVMSupport), we would
attempt to execute `config.guess` to determine the host triple since
`CMAKE_SYSTEM_NAME` is not Windows and `CMAKE_C_COMPILER` will be set to GNU or
Clang.  This will fail as `config.guess` is a shell script which cannot be
executed on Windows.  Simply log a warning instead.  The user can specify the
value for this instead in those cases.

llvm-svn: 363420

llvm-svn: 363419

Summary: Function bodies marked inline in an opencl source are eliminated but MaxBB check may prevent inlining them leaving undefined references.

Reviewers: rampitec, arsenm

Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, Anastasia, t-tye, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D63337

llvm-svn: 363418

[FPEnv] Lower STRICT_FP_EXTEND and STRICT_FP_ROUND nodes in preprocess phase of ISelLowering to mirror non-strict nodes on x86.

I recently discovered a bug on the x86 platform: The fp80 type was not handled well by x86 for constrained floating point nodes, as their regular counterparts are replaced by extending loads and truncating stores during the preprocess phase. Normally, platforms don't have this issue, as they don't typically attempt to perform such legalizations during instruction selection preprocessing. Before this change, strict_fp nodes survived until they were mutated to normal nodes, which happened shortly after preprocessing on other platforms. This modification lowers these nodes at the same phase while properly utilizing the chain.5

Submitted by:	Drew Wock <drew.wock@sas.com>
Reviewed by:	Craig Topper, Kevin P. Neal
Approved by:	Craig Topper
Differential Revision:	https://reviews.llvm.org/D63271

llvm-svn: 363417

Earlier commit has added AMDGPUOperand::isBoolReg(). Turns out
gcc issues warning about unused function since D63204 is not
yet submitted.

Added NFC part of D63204 to have a use of that function and
mute the warning.

llvm-svn: 363416

* Add a common function to setup opt-remarks
* Rename common options to the same names
* Add error types to distinguish between file errors and regex errors

llvm-svn: 363415

Enable 48-bytes of implicit arguments for HIP as well. Earlier it was enabled for OpenCL. This code is specific to AMDGPU target.

Differential Revision: https://reviews.llvm.org/D62244

llvm-svn: 363414

rdar://problem/51464644

Differential Revision: https://reviews.llvm.org/D63311

llvm-svn: 363413

For end-users there is no point in printing dSYM load errors for
system frameworks, since they will all fail and there's nothing they
can do about it. This patch hides them by default and shows them when
--verbose is present.

Differential Revision: https://reviews.llvm.org/D63310

llvm-svn: 363412

As suggested in the post-commit thread for rL363392 - it's
wasteful to have so many runs for larger tests. AVX1/AVX2
is what shows the diff and probably what matters most going
forward.

llvm-svn: 363411

Avoid producing illegal register bank copies for reg_sequence and
phi. The default implementation assumes it is possible to pick any
operand's bank and use that for the result, introducing a copy for
operands with a different bank. This does not check for illegal
copies. It is not legal to introduce a VGPR->SGPR copy, so any VGPR
operand requires the result to be a VGPR.

The changes in getInstrMappingImpl aren't strictly necessary, since
AMDGPU now just bypasses this for reg_sequence/phi. This could be
replaced with an assert in case other targets run into this. It is
currently responsible for producing the error for unsatisfiable
copies, but this will be better served with a verifier check.

For phis, for now assume any undetermined operands must be
VGPRs. Eventually, this needs to be able to defer mapping these
operations. This also does not yet have a way to check for whether the
block is in a divergent region.

llvm-svn: 363410

llvm-svn: 363409

llvm-svn: 363408

Summary:
The IntegerState, and its sepecialization BooleanState, can be used to
simplify the implementation of abstract attributes. The two abstract
state implementations provide storage and helpers to deal with bit-wise
encoded state.

Subscribers: hiraditya, bollu, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D60012

llvm-svn: 363407

I'm not 100% sure about this, since I'm worried about IR transforms
that might end up introducing divergence downstream once replaced with
a constant, but I haven't come up with an example yet.

llvm-svn: 363406

llvm-svn: 363405

This test seems to occasionally fail because editline returns a
different number of lines. Rewrite the message in such a way that we
also see the actual lines when that happens (and not just their count).

Also, clean up the dependencies of the test while I'm in there.

llvm-svn: 363404

This is the family of vector instructions that combine all the lanes
in their input vector(s), and output a value in one or two GPRs.

Differential Revision: https://reviews.llvm.org/D62670

llvm-svn: 363403

llvm-svn: 363402

LLD test case will be fixed in a following commit.

Original commit message:

[yaml2obj] - Allow setting custom section types for implicit sections.

We were hardcoding the final section type for sections that
are usually implicit. The patch fixes that.

This also fixes a few issues in existent test cases and removes
one precompiled object.

Differential revision: https://reviews.llvm.org/D63267

llvm-svn: 363401