lldb recently added a tablegen tool. In order to properly cross compile
lldb standalone there needs to be a mechanism to generate the native
lldb build, analgous to what's done for the NATIVE llvm build. Thus,
we can simply modify this setup to allow for any project to be used.

llvm-svn: 366514

Summary:
LLDB_PATH_TO_{CLANG,LLVM}_BUILD were removed and replaced with
{LLVM,Clang}_DIR. Adjust the NATIVE build to account for this.

Subscribers: mgorny

Differential Revision: https://reviews.llvm.org/D64959

llvm-svn: 366513

This reapplies r366142 with a fix for the failing Windows test.

Original commit message:

Creates universal binary output file from input files. Currently uses
hard coded value for alignment.  Want to get the create functionality
approved before implementing the alignment function.

Patch by Anusha Basana <anusha.basana@gmail.com>

Differential Revision: https://reviews.llvm.org/D64102

llvm-svn: 366512

This will remove the ORCv1 deprecation warnings.

llvm-svn: 366511

r366419 adds nsw to more SCEV expressions, which allows polly to
make more aggressive assumptions about the input expressions.

llvm-svn: 366510

We should re-emit `#pragma once` to ensure the preprocessor will
still honor it when running on minimized sources.

Differential Revision: https://reviews.llvm.org/D64945

llvm-svn: 366509

llvm-svn: 366508

When two .c files define a type of the same name, lldb
just picks one and uses it regardless of context.  That is
not correct.  When stopped in a frame in one of the .c files
that define this type, it should use that local definition.

This commit just adds a test that checks for the correct
behavior.  It is currently xfailed.

llvm-svn: 366507

the xcode project.  This gets it a little closer to
working, but I still have to figure out how to generate
the lldb tablegen backend from the Xcode project.

llvm-svn: 366506

This allows to reduce generated AMDGPUGenAsmWriter.inc by ~100Kb.

Differential Revision: https://reviews.llvm.org/D64952

llvm-svn: 366505

Summary:
This change makes it so that passing --shared-memory is all a user
needs to do to get proper multithreaded code. This default can still
be explicitly overridden for any reason using --passive-segments and
--active-segments.

Reviewers: sbc100, quantum

Subscribers: dschuff, jgravelle-google, aheejin, sunfish, jfb, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D64950

llvm-svn: 366504

Add support for folding G_GEPs into loads of the form

```
ldr reg, [base, off]
```

when possible. This can save an add before the load. Currently, this is only
supported for loads of 64 bits into 64 bit registers.

Add a new addressing mode function, `selectAddrModeRegisterOffset` which
performs this folding when it is profitable.

Also add a test for addressing modes for G_LOAD.

Differential Revision: https://reviews.llvm.org/D64944

llvm-svn: 366503

This is a small extension of !associated, mostly useful for the implementation
convenience of instrumentation passes that RAUW globals with aliases, such
as LowerTypeTests.

Differential Revision: https://reviews.llvm.org/D64951

llvm-svn: 366502

This reverts r366441 (git commit 48104ef7)

This causes clang to fail to compile some file in Skia. Reduction soon.

llvm-svn: 366501

Summary: Fix bug in `wasm-ld`'s `Writer::createInitTLSFunction` that only finds `.tdata` if it's the first section.

Reviewers: tlively, aheejin, sbc100

Reviewed By: sbc100

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

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D64947

llvm-svn: 366500

Summary:
Properly generate the outchain for the `__builtin_wasm_tls_base` intrinsic.

Also marked the intrinsic pure, per @sunfish's suggestion.

Reviewers: tlively, aheejin, sbc100, sunfish

Reviewed By: tlively

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

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D64949

llvm-svn: 366499

Summary:
Other platforms don't have the capability to perform llvm_codesign
step. If LLVM_CODESIGNING_IDENTITY is set then this chunk of code would
attempt to codesign if the target was Apple. But when cross compiling
to Darwin from Linux, for example, this step would fail. So test if the
host is Apple as well.

Subscribers: mgorny, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D64942

llvm-svn: 366498

llvm-svn: 366497

Summary:
Some polish for r365099 which adds a static initializer to
MachOObjectFile. Remove it by moving it to file scope.

Reviewers: smeenai, alexshap, compnerd, mtrent, anushabasana

Reviewed By: smeenai

Subscribers: hiraditya, jkorous, dexonsmith, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D64873

llvm-svn: 366496

llvm-svn: 366495

This causes sections with relative pointers to be marked as read only,
which means that they won't end up sharing pages with writable data.

Differential Revision: https://reviews.llvm.org/D64948

llvm-svn: 366494

The new DriverKit user-land kernel drivers in macOS 10.15 / Catalina 
do not have a main() function or an LC_MAIN load command.  lldb uses
the address of main() as the return address for inferior function
calls; it puts a breakpoint on main, runs the inferior function call,
and when the main() breakpoint is hit, lldb knows unambiguously that
the inferior function call ran to completion - no other function calls
main.

This change hoists the logic for finding the "entry address" from 
ThreadPlanCallFunction to Target.  It changes the logic to first
try to get the entry address from the main executable module,
but if that module does not have one, it will iterate through all
modules looking for an entry address.

The patch also adds code to ObjectFileMachO to use dyld's 
_dyld_start function as an entry address.

<rdar://problem/52343958> 

Differential Revision: https://reviews.llvm.org/D64897

llvm-svn: 366493

llvm-svn: 366492

llvm-svn: 366491

llvm-svn: 366490

Instead of having to write FileSpecList::Append(FileSpec(args)) you can
now call FileSpecList::EmplaceBack(args), similar to
std::vector<>::emplace_back.

llvm-svn: 366489

Add braces around macro `{ MACRO(); }` to guard against macros that
expand to multiple statements.

llvm-svn: 366488

llvm-svn: 366487

While 'd_type' is a non-standard extension to `struct dirent`, only
glibc signals its presence with a macro '_DIRENT_HAVE_D_TYPE'.
However, any platform with 'd_type' also includes a way to convert to
mode_t values using the macro 'DTTOIF', so we can check for that alone
and still be confident that the 'd_type' member exists.

(If this turns out to be wrong, I'll go back and set up an actual
CMake check.)

I couldn't think of how to write a test for this, because I couldn't
think of how to test that a 'stat' call doesn't happen without
controlling the filesystem or intercepting 'stat', and there's no good
cross-platform way to do that that I know of.

Follow-up (almost a year later) to r342089.

rdar://problem/50592673
https://reviews.llvm.org/D64940

llvm-svn: 366486

llvm-svn: 366485

Summary: http://llvm.org/PR39606

Reviewers: Quuxplusone

Subscribers: christof, dexonsmith, libcxx-commits

Differential Revision: https://reviews.llvm.org/D54410

llvm-svn: 366484

If the threadprivate variable is used in the copyin clause on inner
parallel directive with TLS support, we capture this variable in all
outer OpenMP scopes. It leads to the fact that in all scopes we're
working with the original variable, not the threadprivate copies.

llvm-svn: 366483

Incomplete revert. Mea culpa. This test is failing on sanitizer-x86_64-linux
and our internal CI.

llvm-svn: 366482

llvm-svn: 366344 missed an include that broke the LLVM_ENABLE_MODULES
build.

llvm-svn: 366481

The RISC-V hard float calling convention requires the frontend to:

* Detect cases where, once "flattened", a struct can be passed using
int+fp or fp+fp registers under the hard float ABI and coerce to the
appropriate type(s)
* Track usage of GPRs and FPRs in order to gate the above, and to
determine when signext/zeroext attributes must be added to integer
scalars

This patch attempts to do this in compliance with the documented ABI,
and uses ABIArgInfo::CoerceAndExpand in order to do this. @rjmccall, as
author of that code I've tagged you as reviewer for initial feedback on
my usage.

Note that a previous version of the ABI indicated that when passing an
int+fp struct using a GPR+FPR, the int would need to be sign or
zero-extended appropriately. GCC never did this and the ABI was changed,
which makes life easier as ABIArgInfo::CoerceAndExpand can't currently
handle sign/zero-extension attributes.

Re-landed after backing out 366450 due to missed hunks.

Differential Revision: https://reviews.llvm.org/D60456

llvm-svn: 366480

llvm-svn: 366479

r366471 added "-features autodoc" without a trailing comment, leading to `Unrecognized option -features autodoc-threads` due to implicit string concatenation. Add a comma to fix that.

Also separate into "-features" and "autodoc", otherwise it gets parsed as a single "-features autodoc" flag which is also not recognized (it must be two separate CLI args).

llvm-svn: 366478

The IRBuilder doesn't know that FPTrunc and FPExt have constrained
equivalents. Add the support by building on the strict FP mode now
present in the IRBuilder.

Reviewed by:	John McCall
Approved by:	John McCall
Differential Revision:	https://reviews.llvm.org/D64934

llvm-svn: 366477

Summary:
We currently don't support offsetof in the expression evaluator as it is implemented as a macro
(which then calls __builtin_offsetof) in stddef.h. The best solution would be to include that
header (or even better, import Clang's builtin module), but header-parsing and
(cross-platform) importing modules is not ready yet.

Until we get this working with modules I would say we add the macro to our existing macro list
as we already do with other macros from stddef.h/stdint.h. We should be able to drop all of them
once we can import the relevant modules by default.

rdar://26040641

Reviewers: shafik, davide

Reviewed By: davide

Subscribers: clayborg, lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D64917

llvm-svn: 366476

Summary:
Add `__builtin_wasm_tls_base` so that LeakSanitizer can find the thread-local
block and scan through it for memory leaks.

Reviewers: tlively, aheejin, sbc100

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

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D64900

llvm-svn: 366475