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

clang-tidy + clang-format

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

The tests were slightly underspecifying the current behavior of Clang.

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

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.

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

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

These try_lock usages intentionally ignore the result, as commented.

Similar to 4d23764d, which was reviewed
previously.

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

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

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

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.

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

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.

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

This reverts commit 0c67cfdb which has a broken test.

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)

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

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

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

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

This should fix the modules build on the GreenDragon bot.

Re-land "[lldb] [testsuite] generalize `DWARFASTParserClangTests` based on `DWARFExpressionTest`'s YAML"

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

This was broken by 28c91219, which added `struct_stack_t_sz` only to
sanitizer_platform_limits_posix.{cpp,h}.

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

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

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

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.

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

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

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.

Split the VALU 64-bit case in RegBankSelect.