llvm-svn: 360514

No need to select the register class based on type and features. It should
already be setup by X86ISelLowering.

llvm-svn: 360513

We were checking for SSE4.1 for FP types, but not integer 128-bit types.

Fixes PR41837.

llvm-svn: 360512

[X86] Add a test case for idempotent atomic operations with speculative load hardening. Fix an additional issue found by the test.

This test covers the fix from r360475 as well.

llvm-svn: 360511

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360510

Summary:
Breakpoint shouldn't need to depend on any specific details from a
programming language. Currently the only language-specific detail it takes
advantage of are the different qualified names an objective-c method name might
have when adding a name lookup. This is reasonably generalizable.

The current method name I introduced is "GetVariantMethodNames", which I'm not
particularly tied to. If you have a better suggestion, please do let me know.

Reviewers: JDevlieghere, jingham, clayborg

Subscribers: mgorny, lldb-commits

Differential Revision: https://reviews.llvm.org/D61746

llvm-svn: 360509

llvm-svn: 360508

llvm-svn: 360507

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360506

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure

llvm-svn: 360505

We're building with -std=c11 now (as opposed to -std=gnu11), so we can't
use GNU extensions and need to spell inline assembly as __asm__.

llvm-svn: 360503

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360502

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360500

their lifetime in constant expressions.

This is undefined behavior per [class.cdtor]p2.

We continue to allow this for objects whose values are not visible
within the constant evaluation, because there's no way we can tell
whether the access is defined or not, existing code relies on the
ability to make such calls, and every other compiler allows such
calls.

llvm-svn: 360499

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360498

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360497

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360496

The double dash form is documented by GNU as, used by gcc, and accepted by llvm-mc.

llvm-svn: 360495

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360494

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360493

llvm-svn: 360492

llvm-svn: 360491

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360490

llvm-svn: 360489

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360488

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360487

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360486

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360485

For some targets, there is a circular dependency between InstPrinter and
MCTargetDesc.  Merging them together will fix this.  For the other targets,
the merging is to maintain consistency so all targets will have the same
structure.

llvm-svn: 360484

This driver flag is useful when users want to link against the compiler's
builtins, but nothing else, and so use flags like -nostdlib.

Darwin can't use -nolibc & nostdlib++ like other platforms on because we
disable all runtime lib linking with -static, which we still want to have
an option to link with the builtins.

Differential Revision: https://reviews.llvm.org/D58320

llvm-svn: 360483

llvm-svn: 360482

This reverts r360171 (git commit a9d6c32e). A repro showing the asan/msan failures is forthcoming.

llvm-svn: 360481

logging when the testsuite is run with trace mode enabled are leaving
the logging enabled after the tests have finished.  That state
isn't cleared in a --no-multiprocess testsuite run.

llvm-svn: 360480

As requested in D58632, cleanup our red zone detection logic in the X86 backend. The existing X86MachineFunctionInfo flag is used to track whether we *use* the redzone (via a particularly optimization?), but there's no common way to check whether the function *has* a red zone.

I'd appreciate careful review of the uses being updated. I think they are NFC, but a careful eye from someone else would be appreciated.

Differential Revision: https://reviews.llvm.org/D61799

llvm-svn: 360479

llvm-svn: 360478

If a MachO section has the no-dead-strip attribute set then its atoms should
be preserved, regardless of whether they're public or referenced elsewhere in
the object.

llvm-svn: 360477

These yaml test cases appear to have been affected by PR41836

Right now what happens is that these empty .bss sections are merged into
.data, then the .data output section ends up having a zero virtual size,
and it is discarded from the output after addresses are assigned.
However, we've already assigned OutputSections to Chunks, so we don't
correctly report the zero-sized chunks that were in there as having been
discarded. Soon, we will report them as discarded, so these test cases
need to be updated to have a non-zero size so they aren't discarded.

llvm-svn: 360476

After D58632, we can create idempotent atomic operations to the top of stack.
This confused speculative load hardening because it thinks accesses should have
virtual register base except for the cases it already excluded.

This commit adds a new exclusion for this case. I'll try to reduce a test case
for this, but this fix was verified to work by the reporter. This should avoid
needing to revert D58632.

llvm-svn: 360475

necessary.

Prior to r349952, clang used to call objc_msgSend when sending a release
messages, emitting an invoke instruction instead of a call instruction
when it was necessary to catch an exception. That changed in r349952
because runtime function objc_release is called as a nounwind function,
which broke programs that were overriding the dealloc method and
throwing an exception from it. This patch restores the behavior prior to
r349952.

rdar://problem/50253394

Differential Revision: https://reviews.llvm.org/D61803

llvm-svn: 360474

We need to serialize SizeOfRawData through even when there is no data,
as in a .bss section.

Fixes PR41836

llvm-svn: 360473