breaks amdgpu buildbot

When passing arguments with `__fastcall` or `__vectorcall` in 32-bit MSVC, the following arguments have chance to be passed by register if the current one failed. `__regcall` from ICC is on the contrary: https://godbolt.org/z/4MPbzhaMG
All the three calling conversions are not supported in GCC.

Fixes: #57737

Reviewed By: rnk

Differential Revision: https://reviews.llvm.org/D133920

Origin handling is not implemented yet.

Reviewed By: kda

Differential Revision: https://reviews.llvm.org/D133682

Fix error C2027: use of undefined type 'llvm::MemoryBuffer'.

This is for the change of
Differential Revision: https://reviews.llvm.org/D134186

Does not update any of the uses.

Summary: This patch add codegen support for the has_device_addr clause.  It
use the same logic of is_device_ptr.

Differential Revision: https://reviews.llvm.org/D134186

Fixes compile time regression after
bb70b5d4

If we have already calculated the incoming state before, use that
as our starting point to ensure we are conservative.

This fixes an infinite loop found in our downstream where we
we allowed two waves of updates to propagate through a loop and
the merge points allowed us to toggle back and forth between states.
No small reproducer right now.

Differential Revision: https://reviews.llvm.org/D134229

This does not try to pass it through from the end users.

GCC, glibc, binutils, and LLVM have added support for LoongArch64.
This patch adds support for LLVM OpenMP following D59880 for RISCV64.

Reviewed By: MaskRay, SixWeining

Differential Revision: https://reviews.llvm.org/D132925

versions

 It looks like being able to write it was added, but missed the ability
  to read it.

Fixes #57803.

Differential Revision: https://reviews.llvm.org/D134146

Add support for explicitly typed enums:
  enum Foo : unsigned { ... };
to the internal representation and to the YAML output.

Add support for getting the value of an enum constant, as well as accessing the original expression that produced it. This changes the YAML output of enums from an array of strings for the enum members to an array of dictionaries. These dictionaries now report the name, value, and original expression.

The markdown and HTML outputs are unchanged, they still output the name from the new enhanced internal schema.

Reviewed By: paulkirth

Differential Revision: https://reviews.llvm.org/D134055

Reviewed By: aartbik, ThomasRaoux

Differential Revision: https://reviews.llvm.org/D134204

This patch removes the remaining reproducer code. The SBReproducer class
remains for ABI stability but is just an empty shell. This completes the
removal process outlined on the mailing list [1].

[1] https://lists.llvm.org/pipermail/lldb-dev/2021-September/017045.html

This change finalizes the series of patches aiming to replace the old strategy of VGPR to SGPR copy lowering.

  # Following the https://reviews.llvm.org/D128252 and https://reviews.llvm.org/D130367 code parts that are no longer used were removed.
  # The first pass over the MachineFunctoin collects all the necessary information.
  # Lowering is done in 3 phases:
     - VGPR to SGPR copies analysis  lowering
     - REG_SEQUENCE, PHIs, and SGPR to VGPR copies lowering
     - SCC copies lowering is done in a separate pass over the Machine Function

Reviewed By: rampitec

Differential Revision: https://reviews.llvm.org/D131246

We were only setting this flag the first time we added the blocks
not when we mark them for revisiting.

Reviewed By: reames

Differential Revision: https://reviews.llvm.org/D134193

Implement suggested fix for [[ https://cplusplus.github.io/CWG/issues/2628.html | DR2628. ]] Couldn't update the DR docs because there hasn't been a DR index since it was filed, but the tests still run in CI.

Note: I only transfer the constructor constraints, not the struct constraints. I think that's OK because the struct constraints are the same
for all constructors so they don't affect the overload resolution, and if they deduce to something that doesn't pass the constraints
we catch it anyway. So (hopefully) that should be more efficient without sacrificing correctness.

Closes:
https://github.com/llvm/llvm-project/issues/57646
https://github.com/llvm/llvm-project/issues/43829

Reviewed By: erichkeane

Differential Revision: https://reviews.llvm.org/D134145

The compiler does not reorder the gather nodes with reused scalars, just
does it for opernads of the user nodes. This currently does not affect
the compiler but breaks internal logic of the SLP graph. In future, it
is supposed to actually use all nodes instead of just list of operands
and this will affect the vectorization result.
Also, did some early check to avoid complex logic in cost estimation
analysis, should improve compiler time a bit.

This change was originally reviewed as part of what became a4a29438, but apparently forgot to git-add it when applying the patch.  Oops.

    This patch adds a new "target.auto-source-map-relative" setting.

    If enabled, this setting may auto deduce a source map entry based on requested
    breakpoint path and the original path stored in debug info for resolved
    breakpoint.

    As an example, if debug info contains "./a/b/c/main.cpp", user sets a source
    breakpoint at "/root/repo/x/y/z/a/b/c/main.cpp". The breakpoint will resolve
    correctly now with Greg's patch https://reviews.llvm.org/D130401. However, the
    resolved breakpoint will use "./a/b/c/main.cpp" to locate source file during
    stop event which would fail most of the time.

    With the new "target.auto-source-map-relative" setting enabled, a auto deduced
    source map entry "." => "/root/repo/x/y/z" will be added. This new mapping will
    help lldb to map resolved breakpoint path "./a/b/c/main.cpp" back to
    "/root/repo/x/y/z/a/b/c/main.cpp" and locate it on disk.

    If an existing source map entry is used the patch also concatenates the auto
    deduced entry with any stripped reverse mapping prefix (see example below).

    As a second example, debug info contains "./a/b/c/main.cpp" and user sets
    breakpoint at "/root/repo/x/y/z/a/b/c/main.cpp". Let's say there is an existing
    source map entry "." => "/root/repo"; this mapping would strip the prefix out of
    "/root/repo/x/y/z/a/b/c/main.cpp" and use "x/y/z/a/b/c/main.cpp" to resolve
    breakpoint. "target.auto-source-map-relative" setting would auto deduce a new
    potential mapping of "." => "x/y/z", then it detects that there is a stripped
    prefix from reverse mapping and concatenates it as the new mapping:
     "." => "/root/repo/x/y/z" which would correct map "./a/b/c/main.cpp" path to
    new path in disk.

    This patches depends on https://reviews.llvm.org/D130401 to use new added
    SBDebugger::GetSetting() API for testing.

Differential Revision: https://reviews.llvm.org/D133042

In perf2bolt and `-aggregate-only` BOLT mode, the output profile file is written
in fdata format by default. Provide a knob `-profile-format=[fdata,yaml]` to
control the format.
Note that `-w` option still dumps in YAML format.

Reviewed By: #bolt, maksfb

Differential Revision: https://reviews.llvm.org/D133995

Msan has default handler for unknown instructions which
previously applied to these as well. However depending on
mask, not all pointers or passthru part will be used. This
allows other passes to insert undef into sum arguments.
As result,  default strict instruction handler can produce false reports.

Reviewed By: kda, kstoimenov

Differential Revision: https://reviews.llvm.org/D133678

The transformation would fail if none of the sliced dimensions were
linearized by the producing `tensor.collapse_shape`. This is a trivial
edge case but it wasn't correctly tested. Fixes the issue and adds a test.

Reviewed By: nicolasvasilache

Differential Revision: https://reviews.llvm.org/D134088

If libcxxabi is not included CMake will error out:

Cannot find target libcxxabi-SHARED

I ran into this doing the 15.0.0 release

Differential Revision: https://reviews.llvm.org/D133475

Previously when using the `clang-offload-packager` we did not pass the
active offloading kinds. Then in Clang when we attempted to detect when
there was host-offloading action that needed to be embedded in the host
we did not find it. This patch adds the active offloading kinds so we
know when there is input to be embedded.

Reviewed By: tra

Differential Revision: https://reviews.llvm.org/D134189