This reverts commit 51a60ed1, since the test still doesn't pass on
Windows. Marking the test as UNSUPORTED	on Windows again until I've
figured out the problem.

After speaking with Craig Topper about some recent defects, he pointed
out that _ExtInts should be passed indirectly if larger than the largest
int register, and like ints when smaller than that.  This patch
implements that.

Note that this changed the way vaargs worked quite a bit, but they still
work.

Differential Revision: https://reviews.llvm.org/D78785

This matches what is done for MSVC in
b8000c0c. Since that commit, compiler
rt sanitizer libraries aren't linked to with absolute path on windows,
but using their basenames, requiring the libdirs to be passed to
the linker.

This fixes undefined behaviour sanitizer on MinGW after
b8000c0c.

Differential Revision: https://reviews.llvm.org/D79076

The debug directory payload is not located directly after the
debug directory entry itself, but can essentially be located anywhere
in the binary (even outside of mapped sections, although we don't
handle that case).

Differential Revision: https://reviews.llvm.org/D78921

According to the spec, the payload for debug directories can be
in parts of the binary that aren't mapped at runtime - in these
cases, AddressOfRawData is just set to zero.

Differential Revision: https://reviews.llvm.org/D78920

Summary:
Return scalar function when VF==1. The new trivial mapping scalar --> scalar when VF==1 to prevent false positive for "isVectorizable" query.

Author: masoud.ataei (Masoud Ataei)

Reviewers: Whitney (Whitney Tsang), fhahn (Florian Hahn), pjeeva01 (Jeeva P.), fpetrogalli (Francesco Petrogalli), rengolin (Renato Golin)

Reviewed By: fpetrogalli (Francesco Petrogalli)

Subscribers: hiraditya (Aditya Kumar), llvm-commits, LLVM

Tag: LLVM

Differential Revision: https://reviews.llvm.org/D78054

type operands.

The instructions used to convert std.cmpi cannot have i1 types
according to SPIR-V specification. A different set of operations are
specified in the SPIR-V spec for comparing boolean types. Enhance the
StandardToSPIRV lowering to target these instructions when operands to
std.cmpi operation are of i1 type.

Differential Revision: https://reviews.llvm.org/D79049

Instead of completely disabling the tests on Apple, which makes them
disabled on all platforms we test (and hence useless), this commit
disables only the assertions that actually fail. I also created a
bug report to track re-enabling them (https://llvm.org/PR45739).

Removed already-addressed fixme, and updated formatting of a few lines
that were triggering Harbormaster.

These two locale tests are disabled because they were said to "pass in
an uncontrolled manner on Apple platforms". This commit re-enables them
to see what that means, and whether that is still relevant on the
platforms we test.

Before this commit, the tests were either XFAILed or UNSUPPORTED on
Apple and Linux, which is pretty much all the systems we support. If
the tests truly don't work anywhere, they should be removed instead.

Summary:
- Distinguish between partial-profile and non-partial-profile sample PGO.
- Add a flag for partial-profile sample PGO.
- Tune the sample PGO cutoff.
- No default behavior change (yet).

Reviewers: davidxl

Subscribers: eraman, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D78949

This is currently enabled for Intel big cores from Sandy Bridge onward, as well as Atom, Silvermont, and KNL, due to 64-bit division being so slow on these cores. AMD cores can do this in hardware (use 32-bit division based on input operand width), so it's not a win there. But since the majority of x86 CPUs benefit from this optimization, and since the potential upside is significantly greater than the downside, we should enable this for the generic x86-64 target.

Patch By: @atdt

Reviewed By: @craig.topper, @RKSimon

Differential Revision: https://reviews.llvm.org/D75567

Summary:
AArch64's system register ERXTS_EL1 is present in the backend as a
component of the Arm Reliability, Availability and Serviceability (RAS)
extension. However, it has been removed from the specification before
its final release.

This patch removes the register.

Reviewers: SjoerdMeijer, DavidSpickett

Reviewed By: DavidSpickett

Subscribers: DavidSpickett, kristof.beyls, hiraditya, danielkiss, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D79007

Reviewers: dblaikie, craig.topper, wmi

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D79047

The issue we were seeing before should have been fixed by 178a0c80.

On certain targets std.subview should be able to take memrefs from non-zero
addrspaces. Improve lowering logic to llvm dialect and amend the tests.

Differential Revision: https://reviews.llvm.org/D79024

Revert rG5c4b4a62 "PseudoSourceValue.h - reduce GlobalValue.h include to forward declaration. NFC."

Causes buildbot failures.

Enhance lowering logic and tests so vector.transfer_read and
vector.transfer_write take memrefs on non-zero addrspaces.

Differential Revision: https://reviews.llvm.org/D79023

Summary:
Dependencies ensure that Protobufs are generated before all libraries
depending on the headers are **built**, not linked.

Reviewers: sammccall

Reviewed By: sammccall

Subscribers: mgorny, ilya-biryukov, MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D79085

Fix MachineMemOperand.h implicit dependency on Type.h via PseudoSourceValue.h

This reverts commit 03ffe586.

Full tile of reverted commit is:
[ELF][PPC64] Don't perform toc-indirect to toc-relative relaxation for
R_PPC64_TOC16_HA not followed by R_PPC64_TOC16_LO_DS

Breaks the multistage lld PowerPC buildbot.

The cause of this crash is relatively simple -- we are using a
SymbolFileDWARFDwo to parse a (skeleton) dwarf unit. This cause the
CompileUnit to be created with the wrong ID, which later triggers an
assertion in SymbolFile::SetCompileUnitAtIndex. The fix is also simple
-- ensure we use the right symbol file for parsing.

However, a fairly elaborate setup is needed trigger this bug, because
ParseCompileUnit is normally called very early on (and with the right
symbol file object) during the process of accessing a compile unit.
The only way this can be triggered is if the DWARF unit is
"accidentally" pulled into scope during expression evaluation
This can happen if the "this" object used for the context of an
expression is in a namespace, and that namespace is also present in
other compile units

The included test recreates this setup.

Since 88af3ddb, libc++ will prefer Python 3 when available. It is
available on Apple platforms, so subprocess.check_output will return
bytes instead of str. This lead to comparisons against str to be false,
and the MacOS platform not being detected properly.

Summary:
This enforces layering, reduces a sprawling clangd/ directory, and makes life
easier for embedders.

Reviewers: kbobyrev

Subscribers: mgorny, ilya-biryukov, javed.absar, MaskRay, jkorous, arphaman, jfb, kadircet, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D79014

It broke building libclang.dll in Windows builds configured with
LLVM_ENABLE_PIC=OFF. See discussion on the cfe-commits thread.

These are custom expanded in CGBuiltin:

  svmov_b_z(pg, op) <=> svand_b_z(pg, op, op)
  svnot_b_z(pg, op) <=> sveor_b_z(pg, op, pg)

Reviewers: SjoerdMeijer, efriedma, ctetreau, rengolin

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D79039

Move DenseMapInfo specialization to llvm namespace