Test case for failures from
https://lab.llvm.org/buildbot/#/builders/74/builds/26697
caused the revert of 95fef1df in 589c7abb.

This is an exact upstreaming of the downstream diff. Minor
simplifications can be made in the future but upstreaming as-is will
make it easier for us to deal with downstream merge conflicts.

Partially fixes #83805

Hi all,

This patch is a follow-up of #79101. It migrates logic from
`visitVSELECT` to `visitVP_SELECT` to simplify `vp.select`. With this
patch we can do the following combinations:

```
vp.select undef, T, F --> T (if T is a constant), F otherwise
vp.select <condition>, undef, F --> F
vp.select <condition>, T, undef --> T
vp.select false, T, F --> F
vp.select <condition>, T, T --> T
```

I'm a total newbie to llvm and I'm sure there's room for improvements in
this patch. Please let me know if you have any advice. Thank you in
advance!

This is an extension to 07151f02 which handled SandyBridge so we at least model the regression identified in #14640

Confirmed by Agner + uops.info/uica (SkylakeServer also had an incorrect use of Port015 instead of just Port01)

I raised #86669 as a proposal for a 'x86 unfold' pass that can unfold these (if we have the free registers) driven by the scheduler model.

It's a common pattern that we have a machine type, but we don't care
which ARM64* platform we're dealing with. We already have
isAnyArm64 for that, but it does not fit cases where we use a switch
statement. With this helper, it's easy to simplify such cases by using
Triple::ArchType instead of machine type.

This is a proper fix for https://github.com/llvm/llvm-project/issues/40308

Empty structs are ignored for parameter passing purposes, but va_arg was
incrementing the pointer anyway for that the size of empty struct in c++
is 1 byte, which could lead to va_list getting out of sync. Fix #86057.

For pointer types the indices are restricted to one integer-like
operand.
For opaque types no further restrictions are made.

One or two of the tests need slight tweaks to make them pass when
building with musl.

This patch is a re-application of b61fb184 which was reverted in 0847c903
because it broke the build.

rdar://118885724



Co-authored-by: Alastair Houghton <ahoughton@apple.com>

Fixes #84843

FixedVectorType::isValidElementType for consistency.

Reenables b31414bf.

Also adds a new warning for missing `--symbol-graph-dir` arg when
`--emit-extension-symbol-graphs` is provided. This also reverts the
commit that removed.

We were generating "" (the empty string) as an alias for a bunch of FLAT
instructions, which had no effect except to cause tablegen to generate
some very long if-else chains in the generate AsmMatcher.

addMeadata is only over called with a single element, clean up the
variant that takes multiple values.

This was part of https://github.com/llvm/llvm-project/pull/85050.

It is suggested to split the unrelated change as much as possible. So
here is the patch.

[C++20] [Modules] [Driver] Emit unused argument warning if we use '-fmodule-output' with non-module input

We required the file name of an 'importable module unit' should end
with .cppm (or .ccm, .cxxm, .c++m).

But the driver can accept '-fmodule-output' for files with normal
suffixes (e.g., .cpp). This is somewhat inconsistency.

In this patch, we only claim the option `-fmodule-output` is used if
the type of the input file is modules related. Then now the compiler
will emit 'unused argument' warnings if the input file is not modules
related.

This reverts commit 4b25053a.

There're failures in some target.

This tries to fill in the basic vector handling for sadd_sat/uadd_sat
and ssub_sat/usub_sat. It just handles the basics, marking legal types
and clamping illegally sized vectors to legal ones.

We have many problems to use CodeView for a win32-elf target, e.g.,
#87140 and `error: .seh_* directives are not supported on this target`.

Fixes: #87140

Now the `DeclContext::noload_lookup()` asserts that 'this' is not a
transparent context. However, this is not consistent with
`DeclContext::lookup()`, which will lookup into its parent context if
'this' is a transparent context.

This patch makes the behavior of `DeclContext::noload_lookup()` to be
consistent with `DeclContext::lookup()`, to lookup into the parent
context if 'this' is a transparent context.

This patch extract logics in ASTWriter::WriteASTCore about writing decls
and types into a standalone function. The WriteASTCore function is
pretty long and hard to read. It should be helpful for readability to extract the common
logics into a standalone function.

This is also helpful for further changes e.g., removing unreachable
declarations.

expand-large-fp-convert cannot handle vector types.
If overly large vector element types survive into
isel, they will likely be scalarized there, but since
isel cannot handle scalar integer types of that size,
it will assert.

Handle vector types in expand-large-fp-convert by
scalarizing them and then expanding the scalar type
operation. For large vectors, this results in a
*massive* code expansion, but it's better than
asserting.

`and/or/xor` operations can each be changed to sum of logical
operations including operators other than themselves.

 `x&y -> (x|y) ^ (x^y)`
 `x|y -> (x&y) | (x^y)`
 `x^y -> (x|y) ^ (x&y)`

if left of condition of `SelectInst` is `and/or/xor` logical
operation and right is equal to `0, -1`, or a `constant`, and
if `TrueVal` consist of `and/or/xor` logical operation then we
can optimize this case.

This patch implements this combination.

Proof: https://alive2.llvm.org/ce/z/WW8iRR

Fixes https://github.com/llvm/llvm-project/issues/71792.

The check for compressed string is too restrictive (e.g. broke downstream users) and doesn't add much value to the test. Removed it.

Recurrencies in the call graph (even if they are not executed)
prevent computing the minimal stack size required for a kernel
execution. This change disables some functionality of F18 IO
to avoid recursive calls. A couple of functions are rewritten
to work without using recursion.

[Clang] Fix bugs the way we handle duplicate vs conflicting values with loop attribute 'code_align' (#87372)

https://github.com/llvm/llvm-project/pull/70762 added support for new
loop attribute [[clang::code_align()]].

This patch fixes bugs for the test cases below that misses diagnostics due to discontinue to while loop during checking duplicate vs conflicting code_align attribute values in routine CheckForDuplicateLoopAttrs().

[[clang::code_align(4)]]
[[clang::code_align(4)]]
[[clang::code_align(8)]]
for(int I=0; I<128; ++I) { bar(I); }

[[clang::code_align(4)]]
[[clang::code_align(4)]]
[[clang::code_align(8)]]
[[clang::code_align(32)]]
for(int I=0; I<128; ++I) { bar(I); }

Support generating dSYM companion files for (non-Darwin) firmware
environments by considering the binary component of the triple in
addition to the OS component.

rdar://125629792

Before this patch, the size of the reduced BMI may be large than the
full BMI when the source codes is pretty small. This violates the design
principles. The root cause is an oversight that we skipped something
in full BMI but forgot to make it in reduced BMI.

Moves the functions `fsqrt()` and `dsqrt()` from basic functions to
higher math functions in math docs

The unstable partition in partitionRels might reverse IRELATIVE
relocations, so stable_partition in computeRels would lead to IRELATIVE
relocations ordered by decreasing offset. Use stable_partition in
partitionRels to get IRELATIVE relocations ordered by increasing offset.

Re-organizes the tables that listed libc's support for math functions,
and adds two new columns to the tables indicating where the respective
function definitions and error handling methods are located in the C23
standard draft WG14-N3096.

- The opcode of the mina.fmt and max.fmt is documented wrong, the
  object code compiled from the same assembly with LLVM behaves
  differently than one compiled with GCC and Binutils.
- Modify the opcodes to match Binutils. The actual opcodes are as
follows:

  {5,3} | bits {2,0} of func
           |    ...   | 100  | 101    | 110   | 111
  -----+-----+-----+-----+-----+-----
   010  |   ...   |  min  | mina | max  | maxa

This reverts commit 2cf8118e.

Failing tests, revert until I can fix it

Improve the test gnu-ifunc-nonpreemptible.s to check IRELATIVE offsets.
Ensure that IRELATIVE offsets are ordered to improve locality.

[clang][modules] NFCI: Pragma diagnostic mappings: write/read `FileID` instead of `SourceLocation` (#87427)

For pragma diagnostic mappings, we always write/read `SourceLocation`
with offset 0. This is equivalent to just writing a `FileID`, which is
exactly what this patch starts doing.

Originally reviewed here: https://reviews.llvm.org/D137213

Use a single insert for the non-mask case instead of a push_back
followed by an insert that may contain 0 registers.