Summary:
This patch removes the custom CMAKE_RELEASE_CXX_FLAGS variable.
This variable being set was having the effect of removing other important
Release flags, notably `-DNDEBUG`.

This patch may need to be accompanied by fixes for the macOS issues that
the removed comment mentions; I don't have a mac to test this on though so
hopefully a reviewer can help with that.

Reviewers: Andrzej, tskeith, sscalpone

Subscribers: mgorny, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D80794

  * Adds DexLimitSteps Command.
  * Add ConditionalController, a new DebuggerController type.
  * 5 regression tests
  * documentation

  Reviewers: jmorse

  Differential Revision: https://reviews.llvm.org/D79786

  When passing a test path, if the path points directly at a file, then
  normcase would not be called on path.

  This would change the expected lower case drive path, on windows, to be
  uppercase. This patch simply calls normcase on the test path at the earliest
  point possible to avoid this issue.

  Reviewers: djtodoro, jmorse

  Differential Revision: https://reviews.llvm.org/D78633

Reviewers: eduucaldas

Reviewed By: eduucaldas

Subscribers: cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D81009

Summary:
Added initial codegen for 'affinity' clauses on task directives.
Emits next code:
```
kmp_task_affinity_info_t affs[<num_elems>];

void *td = __kmpc_task_alloc(..);

affs[<i>].base = &data_i;
affs[<i>].size = sizeof(data_i);
__kmpc_omp_reg_task_with_affinity(&loc, <gtid>, td, <num_elems>, affs);
```

The result returned by the call of `__kmpc_omp_reg_task_with_affinity`
function is ignored currently sincethe  runtime currently ignores args
and returns 0 uncoditionally.

Reviewers: jdoerfert

Subscribers: yaxunl, guansong, sstefan1, llvm-commits, cfe-commits, caomhin

Tags: #clang, #llvm

Differential Revision: https://reviews.llvm.org/D80240

Summary:
I think we would be better off with tests explicitly specifying the
language mode. Right now Lang_C means C99, but reads as "any C version",
or as "unspecified C version".

I also changed '-std=c++98' to '-std=c++03' because they are aliases (so
there is no difference in practice), because Clang implements C++03
rules in practice, and because 03 makes a nice sortable progression
between 03, 11, 14, 17, 20.

Reviewers: shafik, hlopko

Reviewed By: hlopko

Subscribers: jfb, martong, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D81000

This teaches yaml2obj to allocate file space for a no-bits section
when there is a non-nobits section in the same segment that follows it.

It was discussed in D78005 thread and matches GNU linkers and LLD behavior.

Differential revision: https://reviews.llvm.org/D80629

Instead of using a fake call and metadata to temporarily represent a probed
static alloca, use a pseudo instruction.

This is inspired by the SystemZ approach proposed in https://reviews.llvm.org/D78717.

Differential Revision: https://reviews.llvm.org/D80641

Summary:
For a none-function-like unresolved expression, clang builds a TypoExpr
for it, and tries to correct it afterwards. If the typo-correction
fails, clang just drops the whole expr.

This patch improves the recovery strategy -- if the typo-correction
fails, we preserve the AST by degrading the typo exprs to recovery
exprs.

This would improve toolings for "undef_var" broken cases:
```
void foo();
void test() {
  fo^o(undef_var); // go-to-def, hover still works.
}
```

TESTED=ran tests with this patch + turn-on-recovery-ast patch, it breaks
one declare_variant_messages testcase (the diagnostics are slightly
changed), I think it is acceptable.

```
Error: 'error' diagnostics seen but not expected:
  File llvm-project/clang/test/OpenMP/declare_variant_messages.cpp Line 16: expected 'match' clause on 'omp declare variant' directive
  File llvm-project/clang/test/OpenMP/declare_variant_messages.cpp Line 57: expected 'match' clause on 'omp declare variant' directive
error: 'warning' diagnostics expected but not seen:
  File llvm-project/clang/test/OpenMP/declare_variant_messages.cpp Line 47: the context selector 'kind' in the context set 'device' cannot have a score ('<invalid>'); score ignored
  File llvm-project/clang/test/OpenMP/declare_variant_messages.cpp Line 87: the context selector 'kind' in the context set 'device' cannot have a score ('<invalid>'); score ignored
error: 'warning' diagnostics seen but not expected:
  File llvm-project/clang/test/OpenMP/declare_variant_messages.cpp Line 47: the context selector 'kind' in the context set 'device' cannot have a score ('<recovery-expr>()'); score ignored
  File llvm-project/clang/test/OpenMP/declare_variant_messages.cpp Line 87: the context selector 'kind' in the context set 'device' cannot have a score ('<recovery-expr>()'); score ignored
6 errors generated.
```

Reviewers: sammccall, jdoerfert

Subscribers: sstefan1, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D80733

The pass which infers when it's legal to load a global address space
as SMRD was only considering amdgpu_kernel, and ignoring the shader
entry type calling conventions.

The change to fold_priv_arith looks strange to me, but this was
already the untested behavior for local.

Summary:
ReplayPreamble was just grabbing the reference of IncludeStructure
passed to it and then replayed any includes seen so while exiting
built-in file.

This implies any include seen in built-in files being replayed as part
of preamble, even though they are not. This wasn't an issue until we've
started patching preambles, as includes from built-in files were not
mapped back to main-file.

This patch copies over existing includes at the time of
ReplayPreamble::attach and only replies those to prevent any includes
from the preamble patch getting mixed-in.

Reviewers: sammccall, jkorous

Subscribers: ilya-biryukov, MaskRay, dexonsmith, arphaman, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D80988

This reverts commit bff0c56f. I've
pushed the earlier version of the patch by mistake.

Dialect conversion infrastructure supports 1->N type conversions by requiring
individual conversions to provide facilities to generate operations
retrofitting N values into 1 of the original type when N > 1. This
functionality can also be used to materialize explicit "cast"-like operations,
but it did not support 1->1 type conversions until now. Modify TypeConverter to
support materialization of cast operations for 1-1 conversions.

This also makes materialization specification more extensible following the
same pattern as type conversions. Instead of overloading a virtual function,
users or subclasses of TypeConversion can now register type-specific
materialization callbacks that will be called in order for the given type.

Differential Revision: https://reviews.llvm.org/D79729

ReplayPreamble was just grabbing the reference of IncludeStructure
passed to it and then replayed any includes seen so while exiting
built-in file.

This implies any include seen in built-in files being replayed as part
of preamble, even though they are not. This wasn't an issue until we've
started patching preambles, as includes from built-in files were not
mapped back to main-file.

This patch copies over existing includes at the time of
ReplayPreamble::attach and only replies those to prevent any includes
from the preamble patch getting mixed-in.

The collectCallSiteParameters() method searches for instructions
which load values into registers used for parameters passing.
Previously, interpretation of those values, loaded by one such
instruction, was implemented inside collectCallSiteParameters() method.

This patch moves the interpretation code from collectCallSiteParameters()
method into a separate static method named interpretValue. New method is
called from collectCallSiteParameters() to process each instruction from
targeted instruction scope.

The collectCallSiteParameters() searches for loaded parameter value
among instructions which precede the call instruction, inside the same
basic block. When needed, new method (interpretValue) could be used for
searching any instruction scope.

This is preparation for search of parameter value, loaded inside call
delay slot.

Patch by Nikola Tesic

Differential revision: https://reviews.llvm.org/D78106

One header guard was overlooked when renaming LoopOps to SCF, rename it.
Also drop two unused macros, one of which referred to LoopOp (not "Ops",
hence the overlook).

Summary:
This is a result of the discussion at D78113. Previously we would be
only giving the current offset at which the error was detected. However,
this was phrased somewhat ambiguously (as it could also mean that end of
data was at that offset). The new error message includes the current
offset as well as the extent of the data being read.

I've changed a couple of file-level static functions into private member
functions in order to avoid passing a bunch of new arguments everywhere.

Reviewers: dblaikie, jhenderson

Subscribers: hiraditya, MaskRay, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D78558

Summary:
The code changes are very straight-forward -- just handle both DW_AT_GNU
and DW_AT_call versions of all tags and attributes. There is just one
small gotcha: in the GNU version, DW_AT_low_pc was used both for the
"return pc" and the "call pc" values, depending on whether the tag was
describing a tail call, while the official scheme uses different
attributes for the two things.

Reviewers: vsk, dblaikie

Subscribers: lldb-commits

Tags: #lldb

Differential Revision: https://reviews.llvm.org/D80519

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

Differential Revision: https://reviews.llvm.org/D80523

The lower-to-affine-loops pass in chapters 5-7 of the Toy tutorial has
been creating affine loops, erasing their terminator and creating it
anew using a PatternRewriter instance to work around the fact that
implicit terminators were created without notifying the rewriter. Now
that has been fixed in 3ccf4a5b, remove the code erasing and
re-creating the terminators and rely on the default ones.

This fixes several post-commit nits from D79688 and D80135, namely
typos, debug output and control flow inversion.